JP7603322B2 - Logistics efficiency and optimization evaluation system - Google Patents

Description

This invention relates to a logistics efficiency and optimization evaluation system that uses digital technology to visualize palletizing work, truck loading work, and schedule creation work, which were previously analog (relying on the experience and intuition of on-site workers, etc.), and performs efficiency and optimization verification (including through machine learning), thereby not only eliminating wasteful logistics, but also making logistics look like a "routed bus," and enabling "real-time understanding" of the loading status (of trucks).

Considering the recent efforts in work style reform in the transportation industry, predictions of post-COVID economic recovery, SDGs (Sustainable Development Goals ) , and the movement to reduce _CO2 , it is expected that in the future, about 30% of the current total transportation volume will be difficult to transport. In light of this current situation, not only transportation companies but also customers who request delivery from transportation companies will be forced to think about how to deliver goods efficiently. As a measure, it is desired to be able to build an efficient and optimal logistics system by pre-data management rather than the current performance management (relying on intuition and experience). In other words, it is necessary to collect digital data generated in all logistics processes, visualize it to promote the efficiency of the logistics system, and analyze it to promote the optimization of the logistics system.

Furthermore, the trend towards DX (digital transformation: the concept that the penetration of IT will change people's lives in every aspect for the better) has also reached the logistics industry, and new solutions that have never been seen before will be required to efficiently advance logistics processes. In other words, it is hoped that the efficiency and optimization of logistics will be achieved by controlling the entire logistics system using an IT infrastructure (not just the use of digital data, but also the use of AI such as machine learning and deep learning) for all logistics processes.

Patent Document 1 discloses a method for optimizing logistics through relay points (Patent Document 1: Title of the Invention) that aims to provide a method for optimizing logistics while considering storage constraints at relay points when transporting products from production factories to users. For each product transport volume for a certain period of time for each user, all possible combinations of transport routes for the product are created taking into consideration the factories that can produce the product and the storage capacities at relay points, stock points and deposit points, which depend on the physical properties of the product, and from among these combinations, the transport route that minimizes the total logistics cost, that is, the sum of the user's (transfer costs + loading and unloading costs + storage costs + delivery costs) is calculated (Patent Document 1: Summary).

JP 2002-356232 A

The method for optimizing logistics via relay points according to Patent Document 1 (Patent Document 1: Title of the Invention) requires not only transportation costs but also loading and storage costs in order to transfer products from a factory to a warehouse or temporary storage and deliver them to each user. Considering that the positional relationship between the factory, the warehouse or temporary storage, and the user is the most important factor for reducing logistics costs, the method creates combinations of all transportation routes from the factory of each product to the user via the relay points, corresponding to the storage constraints of the factory and relay points that can produce the product, based on the transportation volume of each product transported to each user, and calculates the transportation route that minimizes the total logistics costs required for transportation from the factory to the user from among these combinations. However, from the perspective of collecting digital data generated in all logistics processes, visualizing and analyzing it, and promoting the use of information to improve the efficiency and optimization of the logistics system, it is important to establish an efficiency and optimization evaluation scheme for the logistics system that pursues efficiency by making the most of the learning model generated by machine learning and performing efficiency and optimization analysis.

The objective of the present invention is to provide a logistics efficiency and optimization evaluation system that uses digital technology to visualize palletizing work, truck loading work, and schedule creation work, which have traditionally been analog (relying on the experience and intuition of on-site workers, etc.), and to perform efficiency and optimization verification (including through machine learning), thereby not only eliminating wasteful logistics, but also making logistics look like a "routed bus," and enabling "real-time understanding" of the loading status (of trucks).

The invention described in claim 1 includes a load aggregation and management platform for inputting and transmitting load data of a plurality of customers, a pre-load data storage means for storing load data inputted and transmitted to the load aggregation and management platform, a palletizing image creation means for creating an image from the load data stored in the pre-load data storage means to instruct a shipping worker to load cargo onto a pallet, an on-board image creation means for creating an image from the load data stored in the pre-load data storage means to instruct a driver to load the pallet onto a truck, a traffic diagram creation means for creating a traffic diagram table and a stowage table from the load data stored in the pre-load data storage means, an efficiency/optimization verification means for verifying whether outputs from the palletizing image creation means, the on-board image creation means, and the traffic diagram creation means are efficient and optimal, an on-board data storage means for storing palletizing image data and on-board image data after efficiency/optimization verification by the efficiency/optimization verification means, and The system is characterized in that it is equipped with an operation data storage means for storing schedule and stowage data, and an optimal inventory calculation means for presenting an optimal inventory quantity to a customer from the palletized image data and on-board image data stored in the on-board data storage means after efficiency and optimization verification by the efficiency and optimization verification means, the schedule and stowage data stored in the operation data storage means after efficiency and optimization verification by the efficiency and optimization verification means, and the load data stored in the advance load data storage means, and the efficiency and optimization verification means verifies delivery confirmation data that is finalized up to two days before the delivery date, and if there are any loads that cannot be loaded onto the vehicle as a result of the verification by the day before the delivery date, it makes an efficiency and optimization proposal to either load them on a later vehicle or have a vehicle with another route stop by and collect the load, and further includes a logistics efficiency and optimization evaluation system equipped with a load shift prevention calculation means using a method using machine learning (generating a trained model by having the system learn ``loading form that prevents load shifting'' and ``delivery route'' as labels (answers) for ``individual forms of loads to be delivered'' and ``delivery location including delivery date and time'').

The invention described in claim 2 is the invention described in claim 1, characterized in that the cargo volume data stored in the advance cargo volume data storage means is the dimensions of the packing boxes forming the luggage, the number of packing boxes (if there are multiple packing boxes, the number of packing boxes for each dimension), the weight of the packing boxes + contents (if there are multiple packing boxes, the weight of the packing boxes + contents for each packing box), and the delivery location (including delivery date and time), which is a logistics efficiency and optimization evaluation system.

The invention described in claim 3 is the invention described in claim 1 or 2, characterized in that the cargo collapse prevention calculation means is a logistics efficiency and optimization evaluation system in which a large amount of data is given with the optimal palletized state image (data is quantified), truck on-board state image (data is quantified), timetable, and stowage table as answers (labels) for the dimensions of the packing boxes forming the cargo, the number of packing boxes (if there are multiple packing boxes, the number of packing boxes for each dimension), the weight of the packing boxes + contents (if there are multiple packing boxes, the weight of the packing boxes + contents for each packing box), and the delivery location (including delivery date and time), and a so-called supervised learned model is used to calculate the optimal palletized state image, truck on-board state image, timetable, and stowage table for the unknown advance cargo amount data.

The invention described in claim 4 is the invention described in claims 1 to 3, characterized in that it is a logistics efficiency and optimization evaluation system equipped with a transport route disclosure means for disclosing the vehicle passing time and vehicle loading status at specified points on a specified travel route by displaying the details as a map on the Internet based on the timetable and loading table created by the operation timetable creation means.

The invention described in claim 5 is the invention described in claim 4, characterized in that it is a logistics efficiency and optimization evaluation system equipped with an optimal route search means for improving the efficiency of the predetermined travel route published by the transportation route publishing means.

This invention relates to a system that obtains digital logistics data (cargo volume data) from customers who request delivery, processes and uses the digital logistics data (cargo volume data) to streamline (including automate) palletizing work, truck loading work, and operation schedule creation work, and evaluates the efficiency of these processes, thereby exploring the efficiency and optimization (including through machine learning) of logistics.

The system includes a load aggregation and management platform for inputting and transmitting load data of multiple customers, a pre-load data storage means for storing the load data inputted and transmitted to the load aggregation and management platform, a palletizing image creation means for creating an image from the data stored in the pre-load data storage means to instruct a shipping worker to load cargo onto a pallet, a vehicle image creation means for creating an image from the data stored in the pre-load data storage means to instruct a driver to load the pallet onto a truck, and a traffic schedule creation means for creating an appropriate schedule table and an appropriate stowage table from the data stored in the pre-load data storage means.

The advance load data storage means for storing data (load data) on packages to be delivered sent from multiple customers can process and store data on packages to be delivered from customers who request delivery (conditions such as package type, delivery date and time, and delivery location: what type of package, where and by when it will be collected, and where and by when it will be delivered, etc.) into digital data in an appropriate format.

The palletizing image creation means creates a visual image of instructions for the shipping worker to load the goods onto the pallet from the advance load data, allowing the shipping worker to visually recognize the state of the goods loaded onto the pallet in an image. The vehicle image creation means creates a visual image of instructions for the driver to load the pallet onto the truck from the advance load data, allowing the shipping worker to visually recognize the state of the pallet loaded onto the truck bed in an image. Therefore, the quality level does not decrease even if the shipping worker or driver changes, and quality can be maintained.

The train schedule creation means creates appropriate schedule tables and appropriate loading tables based on default train schedule data (data for each route and each flight) from advance load data, which eliminates the need for operators to create train schedules based on their intuition and experience. This means that the quality level does not decrease even if the operator changes, and quality can be maintained.

The advance load data is processed and stored in the advance load data storage means by going through the load aggregation and management platform, so that various data formats obtained from customers, such as CSV format, can be converted (including processing) into a format suitable for database storage. Note that it may be possible to consider unifying the data format in the future. Furthermore, the advance load data includes the dimensions of the packing boxes that make up the package, the number of packing boxes (if there are multiple packing boxes, the number of packing boxes for each dimension), the weight of the packing boxes + contents (if there are multiple packing boxes, the weight of the packing boxes + contents for each packing box), and the (respective) delivery location (including the delivery date and time). Note that it goes without saying that other necessary data can be added at any time as necessary.

Furthermore, by using the efficiency/optimization verification means for verifying and optimizing the results obtained from the palletizing image creation means, the on-board image creation means, and the operation schedule creation means, it is now possible to verify the results of comparing the delivery plan data with the delivery performance data, and to make efficiency/optimization proposals to customers after the fact. At the same time, it has become possible to make improvements related to efficiency/optimization within the company. Furthermore, it is now possible to verify delivery confirmation data prior to the delivery date, and to make efficiency proposals to customers by the day before the delivery date.

As an effect of the load shift prevention calculation means (one example of a means for improving efficiency and optimization by machine learning), the optimal loading form that takes into account the delivery destination (delivery order) is acquired by machine learning using a large amount of data, and anyone can reproduce the know-how of loading to prevent "load shifting due to inertia" that occurs when the cargo (including palletized state) loaded on the truck during truck delivery, for example, when the truck brakes suddenly. Therefore, even non-expert workers can perform high-level loading work. To summarize the above, the advantages of the logistics efficiency and optimization evaluation scheme on the company's side are that labor costs can be reduced by reducing the number of people input and that labor shortages can be addressed. The quality level does not decrease even if the shipping worker, driver, or operator is changed, so the work quality can be maintained. By designing logistics in advance, unnecessary truck driving can be reduced, and CO ₂ emissions can be reduced. Furthermore, as advantages for the customer, logistics costs can be reduced and special services (transportation of cargo that could not be accommodated in regular transport) can be suppressed, so the number of vehicles input and transportation design costs can be reduced.

The optimal inventory calculation means presents the optimal inventory quantity to the customer based on the palletizing image data and the on-board image data stored in the on-board data storage means after the efficiency/optimization verification by the efficiency/optimization verification means, the schedule/stowage data stored in the operation data storage means after the efficiency/optimization verification by the efficiency/optimization verification means, and the cargo data stored in the advance cargo data storage means, thereby reducing excess inventory (on the customer's side) and contributing to production management (on the customer's side). (Note that the optimal inventory calculation means can also use the data on the in- and out-of-stock items managed by the so-called "existing in- and out-of-stock management system" before the logistics efficiency and optimization evaluation system according to the present invention is operated.) The transportation company can also save storage space for inventory and allocate that storage space to other operations. In short, waste due to "overproduction" can be reduced. Furthermore, this contributes to reducing _CO2 emissions during production (at the customer's factory) and _CO2 emissions during transportation (on the transportation company's side).

The transport route disclosure means, which displays and publishes the vehicle passing times and vehicle loading status at specified points on a specified driving route in detail as a map on the Internet based on the timetable and loading table created by the operation timetable creation means, provides the customer with useful information (transport route, passing times at each point) when considering the transport method, etc., i.e., by disclosing the transport company's operation status, it has become possible to shorten the time required (for the customer) to consider the transport method, etc. (the time to arrange vehicles for regular services and special services, etc.). For the transport company, by disclosing the transport route, it has become possible to increase the probability that the customer will select it as a transport company and to reduce the time to arrange vehicles when a special service occurs, etc.

By showing an efficient "driving route" by the optimum route search means for making the predetermined driving route published by the transportation route publishing means more efficient, the driving of the vehicle (truck) can be optimized and made more efficient, and waste during driving can be reduced. In other words, not only does it contribute to reducing _CO2 emissions during transportation, but it also leads to reducing the variation in driving time due to the difference between so-called veteran drivers and new drivers. In other words, it has become possible to standardize the quality of drivers.

FIG. 1 is a diagram for explaining an overview of a logistics efficiency and optimization evaluation system according to the present invention. FIG. 13 is a palletizing image diagram created by a palletizing instruction means. FIG. 11 is a truck in-vehicle image diagram created by the in-vehicle instruction means. FIG. 13 is a diagram for explaining method A (post-facto efficiency improvement by comparing planned data with actual data) of the efficiency/optimization verification means. FIG. 13 is a diagram for explaining method B of the efficiency/optimization verification means (proposing efficiency improvements by the day before the delivery date for delivery confirmed data before the delivery date). FIG. 11 is a diagram for explaining an example of a proposal to be made to a customer based on the results obtained from the efficiency improvement/optimization verification means. FIG. 13 is a diagram for explaining a cargo collapse prevention calculation means. FIG. 13 is a diagram for explaining a trained model of the cargo collapse prevention calculation means. FIG. 13 is a diagram for explaining an example of operation by a transportation route disclosure means. FIG. 1 is a diagram for explaining a comprehensive logistics system.

An embodiment of the logistics efficiency and optimization evaluation system 1 according to the present invention will be described in detail below with reference to Figs. 1 to 10. Before describing the details of the logistics efficiency and optimization evaluation system 1, conventional logistics systems and the problems with conventional logistics systems will be described. A transportation company basically goes to its supplier (often the delivery requester: customer) to collect parcels to be delivered (purchasing (collection) work), stores them temporarily in a temporary storage location such as a logistics center, and delivers them to the delivery destination (delivery work). Delivery methods include regular delivery (delivery of parcels to a set location and at a set date and time according to a monthly plan) and special delivery (delivery of parcels that could not be loaded onto a regular delivery by special delivery, etc.).

In the current logistics system, in many cases, operators create operation schedules based on existing operation schedules, taking into account the cargo volume data written on order requests from customers (often paper orders, not digital data), and the operator's intuition and experience. Specifically, operators manually create operation schedules using spreadsheet software, etc., based on cargo volume data from customers (dimensions of each parcel, weight of each parcel, supplier (purchase date and time), delivery destination (delivery date and time), etc. (Note: in this specification, "purchase" and "delivery" are collectively defined as "delivery"). The palletizing work of loading cargo onto pallets is performed by shipping workers based on their intuition and experience, and the loading layout of the palletized pallets onto the truck bed is also considered by the driver based on their intuition and experience. In other words, the palletizing work and the loading work onto the truck bed are performed according to the loading image that the shipping workers and drivers themselves have in their minds.

Because of this reality (the problem was that the packaging was not visual, or in short, not visible), the shipping workers could not properly palletize the packages, or the drivers could not properly load the packages onto the truck bed. In other words, the palletizing and loading of packages onto the truck bed at the shipping site could not be carried out smoothly. This not only significantly extends the shipping time due to the need to redo loading operations, but also leads to some packages that could have been loaded in one go being left behind. In other words, if there is an overflow of packages, a special service (extra service) will be required, resulting in unnecessary costs. Furthermore, when it comes to organizing the operation schedule, there should be an efficient organization skill that cannot be achieved by the operator's intuition and experience alone, that is, an organization skill that avoids the need for special services (extra service) as much as possible.

In order to prevent the occurrence of special (temporary) services in advance, it should be necessary to stop creating operation schedules based on the intuition and experience of operators, and palletizing and truck loading work based on the intuition and experience of shipping workers and drivers, and to make it possible to create appropriate operation schedules even if operators are changed, and to ensure that the quality level of palletizing work does not decrease and quality can be maintained even if shipping workers are changed, and to ensure that the quality level of truck loading work does not decrease and quality can be maintained even if drivers are changed. In the future, taking into account the interests of society as a whole, cargo volume data should be collected as shareable data, and used to the benefit of the logistics industry as a whole.

<Outline of logistics efficiency and optimization evaluation system>
FIG. 1 is a diagram for explaining an outline of the logistics efficiency and optimization evaluation system 1 according to the present invention. As shown in FIG. 1, the logistics efficiency and optimization evaluation system 1 includes a load aggregation and management platform 2 for inputting and transmitting load data of a plurality of customers, a pre-load data storage means 3 for storing the load data inputted and transmitted to the load aggregation and management platform 2, a palletizing image creation means 4 for creating an image to instruct a shipping worker to load a load on a pallet from the data stored in the pre-load data storage means 3, a vehicle image creation means 5 for creating an image to instruct a driver to load the pallet on a truck from the data stored in the pre-load data storage means 3, and a traffic schedule creation means 6 for creating an appropriate schedule table and an appropriate stowage table based on default traffic schedule data (each route and each flight data) from the data stored in the pre-load data storage means 3. The pre-load data may be stored in the pre-load data storage means 3 in a user-friendly processed state via the load aggregation and management platform 2.

It is equipped with an efficiency and optimization verification means 7 for verifying whether the outputs from the palletizing image creation means 4, the on-board image creation means 5, and the operation schedule creation means 6 are efficient and optimal.

The system is equipped with an on-board data storage means 8 that stores palletizing image data and on-board image data after efficiency/optimization verification by the efficiency/optimization verification means 7, and an operation data storage means 9 that stores schedule/stowage data after efficiency/optimization verification by the efficiency/optimization verification means 7. The data stored in the on-board data storage means 8 and the operation data storage means 9 is updated each time the efficiency/optimization verification means 7 performs a verification (of efficiency/optimization).

The system includes an optimal inventory calculation means 11 that calculates the optimal inventory quantity based on the palletizing image data and the on-board image data stored in the on-board data storage means 8 after the efficiency improvement and optimization verification by the efficiency improvement and optimization verification means 7, the schedule and loading data stored in the operation data storage means 9 after the efficiency improvement and optimization verification by the efficiency improvement and optimization verification means 7, and the cargo volume data stored in the advance cargo volume data storage means 3, and presents the optimal inventory quantity to the customer. The system also includes a transport route disclosure means 12 that displays on a map (specifically, on the Internet) the "location" of the "route (including directions)" that the transport company's vehicle (truck whose cargo loading status is also disclosed) will pass (transport route information and real-time transport information), and further includes an optimal route search means 13 that calculates and presents "which road" the transport company's vehicle (truck whose cargo loading status may also be disclosed) should travel on to achieve the most efficient transport. The accuracy of the optimal inventory calculation means 11 and the optimal route search means 13 can be improved by machine learning. The more data is accumulated, the more accurate machine learning can be. The details of the calculation means 10 for preventing cargo from falling during delivery using machine learning techniques (positioned as a means for improving efficiency and optimizing using machine learning techniques) will be described later.

The flow of the logistics efficiency and optimization evaluation system 1 is as follows: based on the advance load data stored in the advance load data storage means 3, which stores the load data input and transmitted to the load aggregation and management platform 2, output is performed by the palletizing image creation means 4, the on-board image creation means 5, and the operation schedule creation means 6. Then, the output by the palletizing image creation means 4, the on-board image creation means 5, and the operation schedule creation means 6 is verified by the efficiency and optimization verification means 7 to verify whether it is efficient and optimal. Then, the (optimized) palletizing image data and on-board image data after the efficiency and optimization calculation by the efficiency and optimization verification means 7 are stored in the on-board data storage means 8, which stores them. At the same time, the (optimized) schedule and stowage data after the efficiency and optimization calculation (by the efficiency and optimization verification means 7) taking into account constraints such as time and place in delivery is stored in the operation data storage means 9, which stores the (optimized) schedule and stowage data.

Even after the palletizing image creation means 4, on-board image creation means 5, and operation schedule creation means 6 have output based on the advance load data stored in the advance load data storage means 3 (if it is before the delivery date), if the data stored in the on-board data storage means 8 that stores the palletizing image data and on-board image data, and/or the data stored in the operation data storage means 9 are updated, the palletizing image data, on-board image data, and schedule/stowage data calculated (verified) using the updated data are often optimal. Therefore, recalculation is performed by the efficiency/optimization verification means 7 based on the updated data, and the palletizing image creation means 4, on-board image creation means 5, and operation schedule creation means 6 are again output (recalculation: see Figure 1).

The optimal inventory quantity calculated by program calculation (including machine learning) from the palletizing image data and on-board image data stored in the on-board data storage means 8 (after efficiency and optimization verification), the schedule and stowage data stored in the operation data storage means 9, and the cargo volume data stored in the advance cargo volume data storage means 3 is presented to the customer (optimal inventory calculation means 11: see Figure 1).

Furthermore, based on the timetable and loading data stored in the operation data storage means 9 (after efficiency and optimization verification), the transportation routes of the transport companies are made public on the Internet (transport route publication means 12 (see Figure 1)), and (taking into account data such as daily traffic conditions) the "roads" on which the transport companies' vehicles (trucks) should travel to achieve the most efficient transportation are calculated and presented, and searched for by the optimal route search means 13 (which may be linked to a global positioning system: GPS) (see Figure 1).

The load aggregation and management platform 2 can be accessed by customers who are delivery requesters to input advance load data, etc. Security measures against intruders are taken so that only those who have been granted access rights, such as customers, can access the platform. For example, it is preferable that security measures such as one-time passwords are taken. The advance load data storage means 3, the in-vehicle data storage means 8, and the operation data storage means 9 are storage media capable of storing (saving) digital data, and any media that can be used as a so-called database can be used, and memory (RAM, etc.) built into a personal computer can be used. Furthermore, they may be composed of non-volatile memory such as flash memory, auxiliary storage devices and media such as hard disks, DVD discs, Blu-ray (registered trademark) discs, cloud storage with security measures, etc.

Figure 2 is a palletizing image created by the palletizing image creation means 4. The shipping worker can perform palletizing work while looking at the screen (or the state printed out on paper) using the palletizing image creation means 4 (see Figure 1). The screen can be displayed in color, so by color-coding, the shipping worker can more clearly recognize (the state when the palletizing work is completed as a picture image). Figure 3 is a truck-mounted image created by the on-board image creation means 5. The driver can perform on-board work in the truck while looking at the screen (or the state printed out on paper) using the on-board image creation means 5 (see Figure 1). The screen can be displayed in color, so by color-coding, the driver can more clearly recognize (the state when the on-board work is completed as a picture image).

The operation schedule creation means 6 is a system that quickly calculates delivery routes, and specifically creates delivery plans, such as "when (what time and minute), which truck will visit which delivery destination (supplier, delivery destination), and in what order," while taking into account complex conditions and constraints such as special circumstances at the delivery site, and designs efficient delivery routes. It creates a schedule table (a diagram showing what time to stop at which base) and a loading table (a table showing which delivery destination's cargo is loaded on each truck). The operation schedule creation means 6 can automatically create operation schedules (by executing a program), so operation schedules can be created without an operator. The operation schedule creation means 6 may be a cloud-based system.

<Measures for efficiency and optimization verification>
Fig. 4 is a diagram for explaining method A (ex-post efficiency improvement by comparing planned data with actual data) of the efficiency/optimization verification means 7. Fig. 5 is a diagram for explaining method B (efficiency proposal by the day before the delivery date for delivery confirmation data before the delivery date) of the efficiency/optimization verification means 7.

Now let us consider Figure 4. A delivery plan for a given month (month N) is (usually) created in month N-1. Normally, the delivery plan data (or cargo volume data) is received from the customer by 4W of month N-1, and the cargo volume aggregation and management platform 2 performs the necessary processing (on the delivery plan data (or cargo volume data) from the customer) and then stores it in the advance cargo volume data storage means 3, thereby finalizing the delivery plan. At this point, it is expected that all deliveries based on the delivery plan created from the delivery plan data (or cargo volume data) from the customer will be completed using only regular routes.

However, in reality, in the actual results for month N, packages that overflowed from regular deliveries were sometimes delivered on additional special deliveries (extra delivery services). Figure 4 shows that three special deliveries (extra delivery services) were made. (In other words, when comparing the plan with the actual results, a discrepancy of three special deliveries occurred.) When the actual data for month N is compiled in the first week of month N+1, a discrepancy between the plan and the actual results (three special deliveries) is seen. In method A of the efficiency and optimization verification means (post-facto efficiency improvement by comparing planned data with actual data), advice is given to the customer regarding orders (efficiency proposals to the customer: see Figure 4) based on verification by comparing the planned data with the actual data (by the efficiency and optimization verification means 7).

Specifically, to avoid generating special services, multiple improvement plans can be presented to the customer, such as loading the cargo loaded on the special service (extra service: incurring extra costs) onto an earlier service (regular service), onto a later service (regular service), onto a service on a different route (existing route), or by rearranging or combining with other routes. If the customer places an order based on these improvement plans, the number of special services can be reduced, i.e., a proposal can be made to review the ordering plan. At this time, a concrete proposal can be made, such as the number of vehicles used could be reduced by "xx units."

Explain about FIG. 5. N day is the delivery date. By N-2 day, the delivery confirmation data to be delivered on N day is verified by the efficiency/optimization verification means 7 based on past data. (The reason why the load data for N day is verified on N-2 day is because the load data provided by the customer one month before the delivery date often fluctuates, and practically speaking, the day when the load data from the customer is confirmed is often 2-3 days before.) If it is found that a special delivery is to be issued as a result, the customer is notified of this and a proposal to change the delivery plan is presented. Specifically, a proposal is made to avoid the occurrence of a special delivery by combining with a previous or subsequent delivery, or by combining with a delivery on a different route. If the verification by the efficiency/optimization verification means 7 shows that not only is it possible to avoid the occurrence of a special delivery (with great effect) but also to reduce regular delivery, a proposal is made to reduce the number of vehicles used for regular delivery. In addition, in order to avoid any disruption to delivery operations, it is desirable to receive final approval from the customer by N-1 day. Promptly thereafter, by the N-1 day, the operation schedule creation means 6 creates a schedule table and a stowage table, and by the Nth day (the day itself), the vehicle image creation means 5 is able to instruct the driver on the optimal vehicle loading state. Needless to say, the palletizing image creation means 4 also instructs the shipping worker on the palletizing work.

Figure 6 is a diagram for explaining examples of proposals made to customers based on the results obtained from the efficiency/optimization verification means 7. As shown in Figure 6, specifically, (1) load the cargo that cannot be loaded onto the vehicle (overflow goods) onto the next vehicle (10 ton truck) to suppress the occurrence of special delivery (optimal candidate A pattern), (2) load the cargo that cannot be loaded onto the vehicle onto a special delivery (4 ton truck (note: lower cost than a 10 ton truck)) (optimal candidate B pattern), (3) stop and collect the cargo that cannot be loaded onto the vehicle on another route vehicle to suppress the occurrence of special delivery (optimal candidate C pattern), etc. can be proposed to the customer. In other words, the reason why multiple optimal delivery patterns are presented and the customer can select one from them according to their convenience is because the customer's interests are also taken into consideration. The present invention is characterized in that it has an effect of reducing costs not only for the company (the entity that uses the efficiency improvement and optimization evaluation scheme of the logistics system according to the present invention) but also for other companies (customers). Furthermore, it goes without saying that the efficiency/optimization verification means 7 can improve (upgrade) the calculation accuracy of the palletizing image creation means 4, the on-board image creation means 5, and the bus schedule creation means 6 by accumulating a large amount of data and using that data for machine learning (supervised learning, unsupervised learning, reinforcement learning, etc.).

<Method to prevent cargo from shifting during delivery using machine learning>
The present invention is characterized in that it does not simply improve the efficiency and optimization of the loading work (taking into consideration the delivery date and time, delivery location (delivery order)) including the palletizing work of the cargo onto the truck and evaluates the degree of completion of the efficiency and optimization, but improves the degree of completion of the efficiency and optimization while learning the optimal loading form by machine learning using a large amount of data to prevent the cargo from collapsing due to inertia, which occurs when the cargo loaded onto the truck suddenly brakes (which may be unavoidable) during delivery. This will be described below. The cargo collapse prevention calculation means 10 is an example of an efficiency and optimization evaluation of logistics using machine learning.

Figure 7 is a diagram for explaining the load collapse prevention calculation means 10. First, the loads are stacked on a flat pallet from the bottom, and then loaded (palletized) on the pallet and then loaded onto the truck bed. When stacking loads on a pallet, the basic rule is to have the heaviest loads at the bottom and the heaviest loads gradually become lighter as they move up. The loads are then loaded onto the truck bed pallet by pallet, and there is a basic rule that the loads are loaded from the very end of the truck bed in the order of the load to be delivered first, the load to be delivered second, the load to be delivered third, and so on (see Figure 7(a)). This is because if the load (pallet) to be unloaded first is not the very end, a space will be created between the loads (pallets) (a gap will be created in the middle), and the load will collapse when the brakes are applied during delivery.

Naturally, the cargo (pallets) is loaded onto the truck when it is stopped (in a very stable state). However, once the truck starts moving, the cargo loaded onto the truck is constantly subjected to external forces such as vibrations and shocks. When the truck starts or stops, centrifugal forces are applied back and forth relative to the direction of travel, up and down when passing over a step, and left and right when turning. Furthermore, the more sudden the action and the greater the speed, the greater the impact. For example, if the cargo loaded onto the truck is in a state like that shown in Figure 7(b), the cargo will collapse due to inertial forces when the truck brakes suddenly. This makes sense in theory, but the current situation is that the know-how for dealing with it has been left to the experience and intuition of each individual driver.

It is difficult to perfectly predict "when, where, at what timing, and how much force will be generated" while driving for a delivery. However, by accumulating data on loading configurations (including the relationship with the individual shapes of the packages to be delivered, the delivery location, and the delivery order) based on the experience and intuition of each driver, and by accumulating a large amount of data on loading configurations (including the individual shapes of the packages to be delivered, the delivery location, and the delivery order) over time, it is believed that a certain correlation will emerge between "the individual shapes of the packages to be delivered, the delivery location, and the delivery order" and "loading configurations that prevent loads from shifting," i.e., a direction for preventing load shifting will become apparent.

Figure 8 is a diagram for explaining the trained model of the cargo shift prevention calculation means 10. A "trained model" is generated by analyzing a large amount of data using machine learning techniques. Specifically, the trained model is generated by learning "loading form that prevents cargo shifting (palletizing image, vehicle image: see Figure 8) and delivery route (descending order obtained from the timetable and stowage table: see Figure 8)" as labels (answers) for "individual forms (dimensions, weight) of cargo to be delivered, and delivery location including delivery date and time (cargo volume data: see Figure 8)."

By learning from large amounts of data, it becomes possible to predict (actually calculate) that "for a specific cargo type (cargo dimensions, weight, etc.) and a specific delivery location (which may include delivery time), this packing form (loading form) and delivery route are optimal (outputting the packing form in the form of an image, and outputting a timetable and stowage table)." In other words, by inputting the new (never before known) "individual cargo type to be delivered and delivery location (which may include delivery time)," the corresponding "optimal loading form (to prevent cargo from falling over), timetable, and stowage table" will be output. In other words, the optimal loading form to prevent cargo from falling over, and the optimal delivery route will be output.

A trained model is generated by providing a large amount of data with the answers (labels) being images of the optimal palletized state (data is quantified), images of the truck on board (data is quantified), a timetable, and a stowage table for the dimensions of the boxes that make up the package (most have set standards, such as 30 cm x 30 cm x 30 cm), the quantity of each box, the weight of each box + contents, and the delivery location (which can also include delivery time), and allowing the model to undergo so-called "supervised learning." The accuracy of machine learning increases with the amount of data, so this process will be carried out continuously into the future. Each time, the trained model is improved, resulting in a trained model with higher accuracy.

As a result of machine learning, when the unknown "dimensions of the boxes that make up the package (many have fixed standards, such as 30 cm x 30 cm x 30 cm), the number of each box, the weight of each box + contents, and the delivery location (which can also include delivery time)" are input, it will not only be possible to efficiently and optimize the loading of packages onto trucks (including palletizing), but it will also be possible to calculate the optimal loading configuration, taking into consideration the prevention of "load collapse due to inertia," which occurs when the package being loaded onto a truck suddenly brakes during delivery. Eventually, even if an inexperienced person joins the company, the AI (artificial intelligence) will be able to show a loading configuration that is equivalent to that of a skilled worker that does not cause load collapse, which will be very favorable.

One loading technique to prevent cargo from shifting is to stack cargo in a different pattern for each tier (alternate row stacking, brick stacking, split stacking, pinwheel stacking, etc. are well known) rather than stacking the cargo in the same order. If the cargo is to be stacked high, it is known that laying interleaf paper in the middle tier will prevent the cargo from sliding sideways, which will also help prevent the cargo from shifting. In addition, cargo-shift prevention wrap, cargo-shift prevention belts, and cargo-shift prevention nets are known to prevent cargo from shifting. It is true that if cargo loaded on a pallet is wrapped in cargo-shift prevention wrap, secured with cargo-shift prevention belts, and covered with cargo-shift prevention nets, it is believed that cargo shifting during delivery can be significantly reduced. However, it is clear that this would result in loading costs that would be too high and would not be profitable.

Experienced shipping workers and drivers often know from past experience the optimal loading configuration to prevent "loads from shifting due to inertia" for each delivery route of each package (they have an image of what they think should be the optimal palletized state and how the packages should be loaded onto the truck), so if this know-how is digitized and input in advance, the time it takes to generate a trained model can be shortened. Even after the trained model is complete, it will continue to be improved as more data is added daily.

When the trained model is improved, the palletizing image data, on-board image data, and schedule/stowage data calculated (estimated) by the improved trained model are optimal, so recalculation is performed based on the improved trained model, and the palletizing image creation means 4, on-board image creation means 5, and schedule creation means 6 output again (recalculation: see FIG. 1). Note that while this specification has mainly described supervised learning as a machine learning method, it goes without saying that in addition to supervised learning, reinforcement learning and unsupervised learning methods can also be adopted as long as they provide the same effects as the present invention.

Figure 9 is a diagram for explaining an example of operation by the transport route disclosure means 12. The transport route disclosure system 12 discloses the transport routes of transport company vehicles (trucks with their cargo loading status also disclosed) and the passing times at each point on the transport route on a special site on the Internet (see Figure 9: Public Map Information (Real-time Transport Information, Transport Route Information)). Member companies (shipper customers or other transport companies) who have the authority to access the special site can plan the transport of products while checking the existing operating routes of the transport companies. After placing an order, customers can check the status of their ordered products (where their products are currently located (on the map) etc.) on their computers (according to the transport plan they have drawn up) (see Figure 9).

Furthermore, the transport route disclosure means 12 contributes greatly to on-site coordination during on-site meetings with customers. That is, when conditions such as transport items and collection/delivery times presented by the customer are entered (on-site) into a tablet terminal or the like during on-site meetings with the customer, artificial intelligence (AI) operates on a system including the transport route disclosure means 12 (the most efficient "travel route" may be selected by the optimal route search system 13), the operation schedule creation means 6, the efficiency/optimization verification means 7, etc., so that an optimal operation schedule (provisional operation schedule) can be automatically created instantly and displayed on the tablet terminal or the like (see FIG. 9). It is also possible to (simultaneously) display an "estimate" on the tablet terminal.

Figure 10 is a diagram for explaining an integrated logistics system (ILS). The integrated logistics system (ILS) is one of the goals of the logistics efficiency and optimization evaluation system according to the present invention. As shown in Figure 10, the integrated logistics system (ILS) is a mechanism in which two or more transport companies (two companies, transport company A and transport company B, are shown in Figure 10, but of course there can be more than that) work together.

Transportation company A and transportation company B each manage data such as cargo volume data, destination platform times, etc., using a system for managing incoming and outgoing goods (incoming and outgoing goods/inventory data management system), a system for managing logistics data (destination condition data management system), etc. (see FIG. 10). As shown in FIG. 10, transportation company A and transportation company B share their data (collecting them into one data group) using the integrated logistics system (ILS), and can obtain information such as optimal timetables and loading tables (timetables and loading data) for joint transportation (with transportation company A and transportation company B as one transportation company community), optimal loading layouts (palletizing image data, and vehicle image data), etc., by consolidating the cargo volume data, destination platform times, etc., managed by the respective systems for managing incoming and outgoing goods and the systems for managing logistics data, etc.

The timetable/stowage data, palletizing image data, and on-board image data are data that have been streamlined and optimized by the efficiency/optimization verification means 7 and then stored by the operation data storage means 9 (where the timetable/stowage data is stored) and the on-board data storage means 8 (where the palletizing image data and on-board image data are stored). Moreover, the data may have had their accuracy improved by machine learning.

(Multiple) member companies (companies that make up the transport company consortium) can not only view information related to their own transport, but also, subject to certain conditions, view information such as optimal timetables and loading tables (timetables and loading data) for joint transport, optimal loading layouts (palletizing image data and on-board image data), etc., and if they determine that it would be more efficient to entrust a transport project to another company rather than to undertake it themselves, they can request transport from each other (see Figure 10).

Furthermore, during meetings with customers (on-site), a "special website" linked to the Integrated Logistics System (ILS) can be referred to (on-site) on a tablet terminal or the like, and the necessary data can be entered to check information such as the optimal timetable for joint transport (with Transport Company A and Transport Company B as one transport company consortium), loading tables (timetable and loading data), and optimal loading layouts (palletized image data and vehicle image data) (however, information from other companies may be subject to conditions such as permission being required) while proceeding with business negotiations. In other words, when data such as quantities related to transport is entered into a tablet terminal or the like (on-site), artificial intelligence (AI) is activated, and timetables, loading tables, etc. are automatically created instantly and can be displayed on the tablet terminal or the like (see Figure 10). In addition, a "quote" can also be displayed on the tablet terminal (at the same time).

<Effects of logistics efficiency and optimization evaluation system>
The logistics efficiency and optimization evaluation system 1 of the present invention obtains cargo volume data in advance from the customer who requests delivery, and uses the cargo volume data (while making full use of AI technology) to make palletizing work, truck loading work, and operation schedule creation work more efficient, and through the above-mentioned series of steps of evaluating the efficiency of these, it aims to improve the efficiency and optimization of logistics, presents the optimal inventory quantity to the customer (contributing to the customer's production management), makes the transportation routes of transportation companies public on the Internet, and allows searches to be made (possibly in conjunction with GPS) to determine which roads the transportation company's trucks (vehicles whose cargo loading status is also disclosed) should travel on to achieve the most efficient transportation, and further achieves the efficiency of logistics by building a system that allows multiple transportation companies to work together.

A major feature of this system is that it has an effect of reducing costs not only for the transport companies that implement the system, but also for the customers who request delivery. By expanding the scope of the system, it is intended to respond to the recent work style reform efforts in the transport industry, the post-COVID economic recovery (forecast), the SDGs (Sustainable Development Goals ) , and the movement toward reducing _CO2 emissions.

Before discussing the efficiency of logistics and the effects of the optimization evaluation system 1, we will discuss the problems that existed when creating a conventional operation schedule. When creating or revising an operation schedule, the daily cargo volume was calculated from the monthly cargo volume (planned base), and the number of trucks to be deployed was set. At this time, trucks were divided into high and low loads based on the full load standard setting per truck (load volume ÷ possible load volume: set arbitrarily).

If the full load standard is set to prevent special (extra) deliveries in anticipation of last-minute additions of freight from customers (known as fluctuating freight volume), the freight will be low-loaded, meaning a large amount of air will be transported, resulting in higher costs. On the other hand, if the full load standard is raised to allow for higher loads, the frequency of special (extra) deliveries will increase if there is a request for additional deliveries, such as an increase in daily production volume, from customers. For this reason, when creating a schedule, the person in charge must check the freight volume on-site and confirm the situation with the customer and the driver, which means that adjustments take a lot of time, and this needed to be improved.

With this invention, by disclosing confirmed daily order data several days prior to the delivery date, it is possible to run an evaluation scheme (Method B of the efficiency and optimization verification means 6 (proposing efficiency improvements by the day before the delivery date for delivery confirmed data prior to the delivery date)) using the efficiency and optimization verification means 6, and it is possible to consider optimal loading and delivery, which makes it possible to prevent the occurrence of special (unscheduled) deliveries.

As a result, the time required for schedule adjustments has been significantly reduced. Clients who request truck-based transport (a freight charge is incurred for each truck) can reduce the number of vehicles used, which should lead to reduced logistics costs and _CO2 emissions. Clients who request individual-based transport (a freight charge is incurred for each package) can reduce the outgoing costs of the receiving party (excessive vehicle input costs, labor costs, etc.) and _CO2 emissions. Furthermore, there was also the issue of frequent changes to the amount of cargo three days or the day before, but now it is possible to make proposals to clients based on clear data, such as "We could have reduced the number of special (extra) flights for this flight."

The benefits for the transport company are that it can increase the loading rate, improve profit margins (individual transport), reduce _CO2 emissions, and reduce the number of special flights (extra service). In addition, it has become possible to reduce the time required for arrangements and the costs incurred for special flights (extra service) at the company's expense.

It is now possible to calculate how to perform palletizing work efficiently, which makes it possible to reduce labor costs by reducing sorting time (improving work efficiency), improve profit margins, and eliminate sorting errors. By giving optimal palletizing instructions, waste is eliminated, and quality does not deteriorate even when workers are changed. It is now possible to calculate how to load items into vehicles, which makes it possible to reduce on-board time, reduce labor costs, improve profit margins, and eliminate on-board errors. Furthermore, by giving optimal on-board instructions, left-behind items and overflows are eliminated, quality does not deteriorate even when drivers are changed, and detailed transportation plans can be made, making it possible to achieve high efficiency from the planning stage. Furthermore, it is now possible to reduce the time required to make transportation plans, which makes it possible to reduce labor costs.

As a benefit to customers, they can increase the loading rate and reduce the number of trips, and they can also reduce special (extra-time) trips as much as possible, which has led to reduced transportation costs. They can simulate how to palletize efficiently, which has reduced sorting time, and they can simulate how to load items efficiently, which has reduced on-board time, which has led to reduced transportation costs.

By outsourcing logistics operations, we can reduce total costs, and by creating detailed transportation plans, we can achieve high efficiency at the planning stage. Furthermore, by shortening the time it takes to create transportation plans, we can reduce labor costs.

The transportation route disclosure means 12 (which may be used in conjunction with the optimum route search means 13) displays the vehicle passing time at specific points on a specific driving route and the vehicle loading status in detail as a map on the Internet, making truck transportation an "omnibus" system in which anyone can see the loading and operation status and operation plans (like a "routed bus" + real-time understanding of the loading status), and by making truck transportation an infrastructure (that is publicly open and available to anyone), the efficiency of transportation itself can be improved.

Specifically, it becomes possible to efficiently select the optimal road (from a number of alternative roads) connecting each of the points that the truck stops at during transportation. The effects of the transportation route disclosure means 12 and the optimal route search means 13 are great, and not only do they contribute to the reduction of _CO2 emissions during transportation in the trucking industry as a whole, but they also lead to the reduction of variations in driving time due to differences between veterans and newcomers, and the standardization of driver quality is achieved, which benefits the entire trucking industry.

When considering the future of the logistics industry, it can be said that the future is far from bright if things continue as they are. In order to overcome this crisis, the logistics industry must work together toward the same goal. For example, it was previously unthinkable to deliver heavy goods (beer, etc.) and light goods (snacks, etc.) mixed together on a truck, but from the perspective of logistics efficiency, mixed loading should be done to reduce transportation costs. Such efforts are beginning to be seen here and there. However, it must be said that there are still high hurdles to overcome before it can be put into practice. By visualizing logistics (visualizing the vehicle loading status, delivery route, etc.) using this invention, it is expected that different industries will come together (including disclosing delivery information to each other) and become a major trend in the logistics industry. In other words, the effect of breaking down the barriers between industries will lead to further development of the logistics industry.

<Examples of changes to the logistics efficiency and optimization evaluation system>
The logistics efficiency and optimization evaluation system according to the present invention is not limited to the aspects of the above-mentioned embodiments, and the configurations of the cargo amount aggregation and management platform, advance cargo amount data storage means, palletizing image creation means, in-vehicle image creation means, operation schedule creation means, efficiency and optimization verification means, in-vehicle data storage means, operation data storage means, cargo collapse prevention calculation means, optimal inventory calculation means, transportation route publication means, optimal route search means, etc. can be appropriately changed as necessary within the scope of the gist of the present invention. For example, the palletizing instruction means and the in-vehicle instruction means can be displayed in a more visual form as a 3D screen.

The logistics efficiency and optimization evaluation system of the present invention has the excellent effects described above, and can therefore be suitably used as a system for improving logistics efficiency by collecting digital data generated in all logistics processes and visualizing it.

1. Logistics efficiency and optimization evaluation system 2. Load aggregation and management platform 3. Advance load data storage means 4. Palletizing image creation means 5. In-vehicle image creation means 6. Traffic schedule creation means 7. Efficiency and optimization verification means 8. In-vehicle data storage means 9. Operation data storage means 10. Load collapse prevention calculation means (an example of an efficiency and optimization means using machine learning)
11. Optimal inventory calculation means 12. Transportation route disclosure means 13. Optimal route search means

Claims (5)

A cargo aggregation and management platform for inputting and transmitting cargo data from multiple customers;
A pre-load data storage means for storing the load data input and transmitted to the load aggregation and management platform;
a palletizing image creating means for creating an image for a shipping worker to instruct how to load the cargo onto a pallet from the cargo weight data stored in the advance cargo weight data storage means;
an on-board image creating means for creating an image based on the load data stored in the advance load data storage means to instruct a driver to load the pallet onto a truck;
a train schedule creating means for creating a train schedule table and a stowage table from the load data stored in the advance load data storage means;
an efficiency/optimization verification means for verifying whether outputs from the palletizing image creation means, the on-board image creation means, and the bus schedule creation means are efficient and optimal;
an on-vehicle data storage means for storing palletizing image data and on-vehicle image data after the efficiency improvement and optimization verification means has verified the efficiency improvement and optimization;
an operation data storage means for storing schedule and loading data after the efficiency and optimization verification by the efficiency and optimization verification means;
an optimal inventory calculation means for presenting an optimal inventory quantity to a customer based on the palletizing image data and the on-board image data stored in the on-board data storage means after the efficiency/optimization verification by the efficiency/optimization verification means, the schedule/loading data stored in the operation data storage means after the efficiency/optimization verification by the efficiency/optimization verification means, and the cargo amount data stored in the advance cargo amount data storage means,
The efficiency/optimization verification means verifies delivery confirmation data that is finalized up to two days before the delivery date, and if there are any packages that cannot be loaded onto the vehicle as a result of the verification up to the day before the delivery date, makes an efficiency/optimization proposal to either load the packages onto a later vehicle or have a vehicle with another route stop by and collect the packages ,
Furthermore, the system is characterized by being equipped with a calculation means for preventing cargo from shifting during delivery, using a method that uses machine learning (a trained model is generated by learning "loading form that prevents cargo from shifting" and "delivery route" as labels (answers) for "individual shapes of cargo to be delivered" and "delivery location including delivery date and time"). The logistics efficiency and optimization evaluation system according to claim 1, characterized in that the cargo volume data stored in the advance cargo volume data storage means is the dimensions of the packing boxes forming the package, the number of packing boxes (if there are multiple packing boxes, the number of packing boxes for each dimension), the weight of the packing boxes + contents (if there are multiple packing boxes, the weight of each packing box + contents), and the delivery location (including delivery date and time). The logistics efficiency and optimization evaluation system according to claim 1 or 2, characterized in that the load collapse prevention calculation means is provided with a large amount of data in which the answers (labels) are an image of the optimal palletized state (data is quantified), an image of the truck loaded state (data is quantified), a timetable, and a stowage table for the dimensions of the packing boxes forming the cargo, the number of packing boxes (if there are multiple packing boxes, the number of packing boxes for each dimension), the weight of the packing boxes + contents (if there are multiple packing boxes, the weight of each packing box + contents), and the delivery location (including delivery date and time), and calculates the optimal image of the palletized state, the image of the truck loaded state, the timetable, and the stowage table for the unknown advance load data using a trained model that has been subjected to so-called supervised learning. The logistics efficiency and optimization evaluation system according to any one of claims 1 to 3, characterized in that it is provided with a transport route disclosure means for disclosing the vehicle passing time and vehicle loading status at specified points on a specified travel route by displaying the details as a map on the Internet based on the timetable and loading table created by the operation timetable creation means. The logistics efficiency and optimization evaluation system described in claim 4 is characterized in that it is provided with an optimal route search means for making the predetermined travel route published by the transportation route publishing means more efficient.

Images