Transportation planning support device and transportation planning support method

The transportation plan creation support device uses adiabatic quantum computation and the Ising model to efficiently create optimal delivery routes that adhere to multiple constraints, addressing the challenges of complex supply chains and reducing computational demands.

JP7591007B2Active Publication Date: 2024-11-27HITACHI LTD
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
JP2022120950
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2022-07-28
Publication Date
2024-11-27
Estimated Expiration
2042-07-28

AI Technical Summary

Technical Problem

Existing methods for creating transportation plans in complex supply chains face challenges due to redundant constraints and the need for extensive computational resources, making it difficult to balance time and plan validity, especially when using quantum computers for annealing methods.

Method used

A transportation plan creation support device utilizing adiabatic quantum computation and the Ising model, where vehicle delivery routes are represented as spins, allowing for efficient calculation of optimal delivery routes that adhere to multiple constraints and evaluation criteria using an Ising model.

Benefits of technology

Enables the prompt creation of transportation plans that comply with numerous restrictions while optimizing evaluation criteria, reducing computational time and resource requirements, and ensuring the validity of the plan.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007591007000016
    Figure 0007591007000016
  • Figure 0007591007000017
    Figure 0007591007000017
  • Figure 0007591007000018
    Figure 0007591007000018
Patent Text Reader

Abstract

To quickly formulate a transportation plan having many constraints in conformity with the constrains.SOLUTION: In a transportation plan formulation support system, a transportation plan formulation support apparatus 100 which can communicate with a plurality of user terminals is configured to: store first constraint conditions from which the deviation is not allowed and second constraint conditions which are related to each evaluation values and from which the deviation is allowed; generate combinations of routes of delivery of each vehicle, as delivery route candidates, for each of evaluation criteria, so as to satisfy the first constraint conditions and optimize the evaluation values based on the evaluation criteria; calculate an Ising model 400 in which adoption of the routes of each delivery of each vehicle is used as spins and sensitivity between variables in a constraint condition function is set as strength of interaction between the spins, regarding an objective function that includes evaluation values of the delivery route candidates for each of the evaluation criteria and the constraint condition function which is minimum when the second constraint conditions are met as terms; and output information indicating adoption of the routes of the vehicles to an output device.SELECTED DRAWING: Figure 2
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a transportation plan creation support device and a transportation plan creation support method. [Background technology]

[0002] In supply chains where transportation operations have become more complicated due to globalization, efficiency has become an increasingly important issue. Mathematical optimization of transportation operations in supply chains is often performed by solving optimization problems based on mathematical programming, which searches for a solution that maximizes or minimizes desired parameters under given conditions.

[0003] As a solution method for mathematical programming, for example, Patent Document 1 discloses a mathematical programming calculation device. Specifically, it is described that when a mathematical programming problem is input, a solution search policy optimization means uses a genetic algorithm to generate individuals having chromosomes indicating a solution search policy, and a population is formed from these individuals. Then, this population is passed to a solution search means, which searches for a solution to the mathematical programming problem according to the solution search policy indicated by the chromosome of each individual in the population, creates solution candidates, and optimizes the solution search policy with the solution candidates as a solution candidate group, and the solution search policy optimization means calculates a fitness value of each solution candidate in the solution candidate group, and if there is no need to search for solution candidates of the next generation, the solution candidate with the best fitness value is set as the solution to this mathematical programming problem.

[0004] Patent document 2 also describes that a transportation planning device holds a number of orders, each of which includes delivery work and collection work, order information indicating the bases where the delivery work for each of the multiple orders is performed and the bases where the collection work is performed, and work time information indicating the work times for each of the delivery work, collection work, and delivery and collection work in which the delivery work and the collection work are considered as one work, generates a transportation work sequence plan indicating the order in which the delivery work and collection work included in the multiple orders are performed, considers consecutive delivery work and collection work at the same base as delivery and collection work, and identifies the work times for the delivery work, collection work, and delivery and collection work included in the transportation work sequence plan by referring to the work time information.

[0005] On the other hand, when optimizing a supply chain, the number of candidate solutions can become explosively large. In such cases, it can be difficult to solve the problem within a practical time frame unless a computer with appropriate computing power, such as a supercomputer or quantum computer, is used.

[0006] Conventional technologies related to quantum computers include, for example, a technology proposed for a computer that enables high-speed calculations for inverse problems or combinatorial optimization problems that require exhaustive search, in which spin is used as a variable in the calculation, the problem to be solved is set in terms of spin-spin interactions and local fields acting on each spin, and all spins are oriented in one direction by an external magnetic field at time t = 0, the external magnetic field is gradually reduced so that the external magnetic field becomes zero at time t = τ, and each spin is assumed to have its orientation determined by the effective magnetic field determined by the combined effects of the external magnetic field and spin-spin interactions at each site at time t as it evolves over time, and in this case, the spin orientation is not completely aligned with the effective magnetic field, but is instead quantum mechanically corrected, so that the system maintains almost the ground state (see Patent Document 3).

[0007] Patent Document 4 discloses a method for creating a delivery plan using an annealing machine, which is a type of quantum computer, including the steps of creating multiple delivery patterns, inputting the multiple delivery patterns into the annealing machine and narrowing down the multiple delivery patterns to a specified number of delivery patterns based on multiple solutions calculated, and creating a delivery plan based on the delivery patterns selected from the specified number of delivery patterns. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Japanese Patent Application Publication No. 10-055439 [Patent Document 2] JP 2020-181282 A [Patent Document 3] International Publication No. 2016 / 157333 [Patent Document 4] JP 2021-111276 A Summary of the Invention [Problem to be solved by the invention]

[0009] However, when applying annealing techniques to a global supply chain where there are a wide variety of transportation routes and constraints related to that transportation, the formula to be formulated becomes redundant, and there is a high possibility that the constraints to be observed will be violated. On the other hand, it is expected that methods such as mathematical programming will take an enormous amount of time to calculate. As such, as supply chains become more complex, there is a problem in that it becomes more difficult to balance the time required to create a transportation plan with the validity of the plan created.

[0010] The present invention has been made in consideration of the above circumstances, and has an object to provide a transportation plan creation support device and a transportation plan creation support method that are capable of supporting the rapid creation of transportation plans having many constraints while complying with those constraints in accordance with their content. [Means for solving the problem]

[0011] One of the present inventions for solving the above problem is a device for supporting the creation of a transportation plan in which a plurality of vehicles deliver each package, the transportation plan creation support device comprising: a storage unit that stores a first constraint condition that is a constraint condition for the delivery and from which deviation is not permitted; and a second constraint condition that is a constraint condition for each evaluation value based on each of a plurality of evaluation criteria for the delivery and from which deviation is permitted; a delivery route candidate creation process that creates, for each of the plurality of evaluation criteria, a combination of delivery routes for each of the vehicles as a delivery route candidate such that an evaluation value based on the evaluation criteria is optimized while satisfying the first constraint condition; an optimal delivery route creation process that calculates an Ising model in which the use of each delivery route for each of the vehicles is set as a spin and the sensitivity between variables in the constraint function is set as the strength of interaction between the spins, for an objective function including, as a term, the evaluation value of the delivery route candidate for each of the evaluation criteria and a constraint function that is minimized when the second constraint condition is satisfied; and a result output process that outputs information indicating the use of each route for each of the vehicles to an output device based on the result of the calculation. Effect of the Invention

[0012] According to the present invention, it is possible to support the rapid creation of a transportation plan having many constraints while observing the constraints according to their contents. Configurations and effects other than those described above will become apparent from the following description of the embodiments. [Brief description of the drawings]

[0013] [Figure 1] 1 is a diagram illustrating an example of a configuration of a transportation plan creation support system according to an embodiment of the present invention. [Diagram 2] 1 is a diagram illustrating an example of hardware and functions included in a transportation plan creation support device according to an embodiment of the present invention. [Diagram 3] FIG. 4 is a diagram illustrating an example of a data configuration of a vehicle information table. [Figure 4] FIG. 13 is a diagram illustrating an example of a data configuration of a base information table. [Diagram 5] FIG. 13 is a diagram illustrating an example of a data configuration of an order information table. [Figure 6] FIG. 11 is a diagram illustrating an example of a data configuration of a travel time information table. [Figure 7] FIG. 13 is a diagram illustrating an example of a data configuration of a cost information table. [Figure 8] FIG. 13 is a diagram illustrating an example of a data configuration of an evaluation criteria information table; [Figure 9] FIG. 13 is a timing chart for obtaining the ground state of a spin system. [Figure 10] FIG. 11 is a flow diagram illustrating an example of a solution in an Ising model. [Figure 11] FIG. 11 is a flow diagram illustrating an example of a transportation plan creation support process. [Figure 12] FIG. 11 is a flow diagram showing an example of a process for determining a delivery route candidate that satisfies a first constraint condition and optimizes delivery costs when the evaluation criterion is delivery costs. [Figure 13] FIG. 13 is a diagram showing examples of delivery route candidates. [Figure 14] FIG. 11 is a flow diagram illustrating details of an optimum delivery route creation process. [Figure 15] FIG. 13 is a diagram showing an example of data acquired by an optimal delivery route creation unit; [Figure 16] FIG. 1 is a diagram illustrating an example of an optimal delivery route. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] <About the annealing machine> As described in the above-mentioned Patent Document 3, the applicant has developed quantum computing technology and has attempted to solve various problems in, for example, exhaustive search problems based on big data (including the concept of combinatorial optimization problems).

[0015] Quantum computers are generally expected to solve such exhaustive search problems. Quantum computers are made up of basic elements called quantum bits, which can simultaneously realize "0" and "1". As a result, all solution candidates can be calculated simultaneously using the initial value, and there is a possibility that exhaustive search can be realized. However, quantum computers need to maintain quantum coherence throughout the entire calculation time.

[0016] One method that has been attracting attention in this context is called adiabatic quantum computing (Reference: E. Farhi, et al., "A quantum adiabatic evolution al gor ithm applied to random instances of an NP-complete problem," Science292, 472 (2001).). This method transforms the problem so that the ground state of a certain physical system is the solution, and attempts to obtain the solution by finding the ground state.

[0017] Let H^p be the Hamiltonian of the physical system for which the problem is set. However, at the start of the calculation, the Hamiltonian is not H^p, but a different Hamiltonian, H^0, whose ground state is clear and easy to prepare, is used. Next, a sufficient amount of time is taken to transition the Hamiltonian from H^0 to H^p. If a sufficient amount of time is taken, the system will remain in the ground state, and the ground state of the Hamiltonian H^p will be obtained. This is the principle of adiabatic quantum computing. If the calculation time is τ, the Hamiltonian becomes equation (1), [Formula 1] JPEG0007591007000001.jpg38170The solution is obtained by evolving the time based on the Schrödinger equation (2).

[0018] [Formula 2] JPEG0007591007000002.jpg30170Adiabatic quantum computing can be applied to problems that require exhaustive search, and a solution is reached in a one-way process. However, if the computation process needs to follow the Schrödinger equation in equation (2), quantum coherence must be maintained, just like in a quantum computer.

[0019] However, while a quantum computer repeats gate operations between one or two quantum bits, adiabatic quantum computing involves simultaneous interaction across the entire quantum bit system, and the concept of coherence is different.

[0020] For example, consider a gate operation on a certain quantum bit. In this case, if there is an interaction between that quantum bit and other quantum bits, it will cause decoherence, but in adiabatic quantum computing, all quantum bits interact at the same time, so decoherence does not occur in this example. Reflecting this difference, adiabatic quantum computing is considered to be more robust against decoherence than quantum computers.

[0021] As described above, adiabatic quantum computing is effective for solving difficult problems that require exhaustive search. The problem to be solved is set up in terms of spin-spin interactions and local fields acting on each spin.

[0022] At time t = 0, all spins are oriented in one direction by an external magnetic field, and the external magnetic field is gradually reduced so that it becomes zero at time t = τ.

[0023] Each spin evolves over time, assuming that its orientation is determined by the effective magnetic field at time t, which is determined by the combined effects of the external magnetic field at each site and the spin-spin interactions.

[0024] In this case, the spin direction is not perfectly aligned with the effective magnetic field, but is quantum mechanically corrected so that the system maintains approximately the ground state.

[0025] In addition, a term (relaxation term) that maintains each spin in its original orientation during time evolution is added to the effective magnetic field, improving the convergence of the solution. Based on the above, an embodiment of the present invention will be described in detail with reference to the drawings.

[0026] <Transportation planning support system> 1 is a diagram showing an example of the configuration of a transportation plan creation support system 1 according to this embodiment. As shown in the figure, the transportation plan creation support system 1 includes a transportation plan creation support device 100 and one or more user terminals 200.

[0027] The transportation plan creation support device 100 and each user terminal 200 are connected via a wired connection such as the Internet, a LAN (Local Area Network), a WAN (Wide Area Network), or a dedicated line. Alternatively, communication is possible via a wireless communication network 10.

[0028] The transportation plan creation support device 100 is a system for a transportation company that receives many orders for delivery of specific cargo from a specific departure point (collection point) to a destination point (unloading point). It is an information processing device that supports the creation of transportation plans for delivery vehicles (such as trucks) to carry out their delivery tasks while satisfying various constraints.

[0029] In this embodiment, the transportation plan creation support device 100 is configured to include an annealing machine that performs adiabatic quantum computing as described below, but of course, the present invention is not limited to this, and any device that can appropriately solve a combinatorial optimization problem in accordance with the transportation plan creation support method of the present invention is applicable. Note that the outline of the annealing machine has already been described based on Patent Document 3, and details of its specific configuration, operation, etc. will be omitted as appropriate (the same applies below).

[0030] In addition, the transportation plan creation support device 100 uses electronic circuits (digital It may be implemented not only in hardware (such as a circuit) but also in a superconducting circuit. The transportation plan creation support device 100 may also be equipped with hardware that realizes an Ising model using a method other than the annealing method. Such methods include, for example, a laser network method (optical parametric oscillation) or a quantum neural network. As described above, although some of the concepts are different, the present invention can also be applied to a quantum gate method in which the calculations performed in the Ising model are replaced with gates such as a Hadamard gate, a rotation gate, or a controlled NOT gate. It is possible to achieve this.

[0031] Next, the user terminal 200 is an information processing device that requests the transportation plan creation support device 100 to create a transportation plan and displays information on the transportation plan created by the transportation plan creation support device 100.

[0032] The user of this user terminal 200 is, for example, a transport company that manages the transport of luggage using multiple vehicles. Each vehicle may deliver multiple luggage at once by passing through multiple collection and unloading bases from its initial departure point to its final destination.

[0033] <Transportation planning support device> 2 is a diagram showing an example of the hardware and functions of the transportation plan creation support device 100 of this embodiment. The user terminal 200 also has similar hardware.

[0034] First, the transportation plan creation support device 100 includes the following hardware components: a storage unit 101 , a memory 103 , a calculation unit 104 , and a communication unit 105 .

[0035] The storage unit 101 may be a suitable storage device such as a solid state drive (SSD) or a hard disk drive. It is composed of non-volatile memory elements.

[0036] The memory 103 is composed of a volatile storage element such as a RAM (Random Access Memory).

[0037] The calculation unit 104 reads out the program 102 stored in the storage unit 101 into the memory 103 and executes the program to overall control the device itself and to perform various determinations, calculations, and control processes. The calculation unit 104 is, for example, a CPU (Central Processing Unit), a DSP (Digital Signal Processor), a GPU (Graphics Processing Unit), or an FPGA (Field-Programmable Gate Array).

[0038] The communication unit 105 is connected to the communication network 10 and is responsible for communication processing with the user terminal 200. The communication unit 105 is, for example, a network interface card (NIC), a wireless communication module, a universal serial interface (USB) module, or a serial communication module.

[0039] The transportation plan creation support device 100 may further include an input device that accepts key input or voice input from a user, or an output device such as a display that displays processed data.

[0040] Next, functions of the transportation plan creation support device 100 will be described.

[0041] As shown in FIG. 2, the transportation plan creation support device 100 includes a delivery route candidate creation unit 111, an optimal delivery route creation unit 112, and a result output unit 113.

[0042] The delivery route candidate creation unit 111 creates a combination of delivery routes for each vehicle that satisfies a constraint condition (hereinafter referred to as a first constraint condition) that is a constraint condition for delivery and from which deviation is not permitted, and maximizes an evaluation value based on a predetermined evaluation criterion, as a delivery route candidate for each of a plurality of evaluation criteria. The details of the evaluation criterion will be described later.

[0043] The optimal delivery route creation unit 112 calculates an Ising model 400 in which the use (adoption) of each delivery route for each vehicle is set as a spin, and the sensitivity between variables in the constraint function is set as the strength of interaction between spins, for an objective function including, as terms, the evaluation value of the delivery route candidate for each evaluation criterion and a constraint function for which deviation is permitted and which is a constraint condition related to each evaluation value based on each evaluation criterion (hereinafter referred to as a second constraint). Details of the Ising model 400 will be described later.

[0044] The result output unit 113 outputs information indicating whether each route is used for each vehicle (i.e., information regarding the transportation plan) to an output device of the transportation plan creation support device 100 or the user terminal 200 based on the calculation result of the optimal delivery route creation unit 112.

[0045] Next, the transportation plan creation support device 100 stores the following tables (databases): a vehicle information table 310, a base information table 320, an order information table 330, a travel time information table 340, a cost information table 350, an evaluation criterion information table 360, and a delivery route candidate information table 370.

[0046] The vehicle information table 310 stores information on each vehicle performing a transportation operation and constraints related to the delivery burden imposed by each vehicle.

[0047] The base information table 320 stores information on bases where each vehicle collects and unloads cargo in a transportation operation and on constraints at each base. The base information table 320 is associated with map information (not shown), and the transportation plan creation support device 100 can calculate the distance between bases (vehicle travel distance).

[0048] The order information table 330 stores information about each order that is requested to be fulfilled. In this embodiment, an order is the delivery of a certain amount of cargo from a collection point to a discharge point.

[0049] The travel time information table 340 stores information regarding travel times between base stations by vehicle.

[0050] The cost information table 350 stores information regarding costs relating to delivery.

[0051] The evaluation criteria information table 360 ​​stores information on the evaluation criteria related to delivery of each vehicle and the method of calculating the evaluation value.

[0052] The delivery route candidate information table 370 stores information such as delivery route candidates calculated by the optimum delivery route creation unit 112 .

[0053] Next, each table will be described in detail.

[0054] (Vehicle information table) 3 is a diagram showing an example of the data configuration of the vehicle information table 310. The vehicle information table 310 has data on a vehicle ID 311 of each vehicle, a maximum operating time 312 of each vehicle (maximum time during which delivery work can be performed), a maximum operating distance 313 of each vehicle (maximum distance that delivery can be performed), a maximum cargo capacity loading capacity 314 of each vehicle (limit on cargo capacity), and a maximum cargo weight loading capacity 315 of each vehicle (limit on cargo weight).

[0055] (Base information table) 4 is a diagram showing an example of the data configuration of the base information table 320. The base information table 320 has data on base ID 321 of each base, operating time period 322 of each base, position 324 of each base (position information, etc., used for calculating the distance between bases, etc.), and enterable vehicles 325, which are types of vehicles that can enter each base.

[0056] (Order information table) 5 is a diagram showing an example of the data configuration of the order information table 330. The order information table 330 has data on an order ID 331 of each order, the amount of luggage to be delivered for each order (amount of luggage 332), a collection base 333 for the luggage for each order, a base 334 for unloading the luggage for each order, a collection time period 335 at the collection base, and a time period 336 when unloading is possible at the unloading base.

[0057] In this way, constraints related to vehicles, packages, or delivery burdens are set in the vehicle information table 310, base information table 320, and order information table 330. In other words, these constraints are first constraints, and are strict constraints from which deviation is not permitted in the delivery of packages.

[0058] (Travel time information table) 6 is a diagram showing an example of the data configuration of the travel time information table 340. The travel time information table 340 has data on a delivery origin base 341 for each delivery, a delivery destination base 342 for each delivery, and a vehicle travel time 343 for each delivery.

[0059] (Cost information table) 7 is a diagram showing an example of the data configuration of the cost information table 350. The cost information table 350 has data on a vehicle ID 351 of each vehicle and a unit cost 352 of the vehicle (e.g., a delivery cost per unit time, a delivery cost per unit travel distance).

[0060] (Evaluation criteria information table) 8 is a diagram showing an example of the data configuration of the evaluation criterion information table 360. The evaluation criterion information table 360 ​​has data on each evaluation criterion 361 and a calculation method 362 of an evaluation value for each evaluation criterion.

[0061] In the example shown in the figure, the evaluation criteria 361 include an evaluation criterion related to the load of luggage, an evaluation criterion related to the vehicle's operating time, an evaluation criterion related to the vehicle's travel distance, and an evaluation criterion related to the vehicle's delivery cost. In this embodiment, the calculation method 362 for the evaluation values ​​related to these evaluation criteria is an arbitrary formula in which the larger the calculated evaluation value, the better the evaluation. For example, For example, a calculation formula such that the greater the load of luggage, the higher the evaluation value, the shorter the vehicle's operating time, the higher the evaluation value, the shorter the travel distance of the vehicle, and the lower the vehicle cost (cost based on time or cost based on distance) is, is set in the evaluation criterion information table 360. The vehicle cost is calculated based on, for example, the base information table 320, the travel time information table 340, and the cost information table 350.

[0062] The evaluation criteria and the calculation method of the evaluation value described here are merely examples, and for example, the number of orders processed by a vehicle in all deliveries may be used as the evaluation criterion (the calculation method).

[0063] 2, the transportation plan creation support device 100 stores an Ising model 400 that employs adiabatic quantum computing. The Ising model 400 will be described in detail later. The Ising model 400 is set in advance by an administrator of the transportation plan creation support device 100, etc.

[0064] The adiabatic quantum computing described in the overview of the annealing machine is also called quantum annealing, and is an extension of the concept of classical annealing to quantum mechanics. That is, adiabatic quantum computing is originally capable of classical operation, and can be interpreted as adding quantum mechanical effects to improve performance in terms of speed and accuracy rate of solutions. Therefore, in this embodiment, the calculation unit 104 of the transportation plan creation support device 100 itself is classical, and a parameter determined by quantum mechanics is introduced into the calculation process to realize a calculation method and device that is classical but includes quantum mechanical effects, and performs calculations using the Ising model 400. However, it is of course possible to adopt a form in which the calculation unit 104 is configured by a quantum computer.

[0065] Based on the above, the following describes the classical algorithm by which the transportation plan creation support device 100 obtains the ground state as a solution of the Ising model 400, and the configuration of the Ising model 400 for realizing this algorithm, while explaining the relationship with adiabatic quantum computing.

[0066] The transportation plan creation support device 100 has N variables sj z (j=1,2,…,N) -1≦sj z ≦1, and set up an Ising model with local fields gj and variable interactions Jij (i, j = 1, 2, ..., N).

[0067] Specifically, the calculation unit 104 divides the time into m parts and performs calculations from t=tk (tk=0) to tm (tm=τ), and calculates the variable Sj at each time tk. Z When calculating (tk), the variable Sj at the previous time tk-1 z Using the value of (tk-1)(i=1,2,..,N) and the relaxation term coefficient 9pina or 9pinb, Bj z (tk) = {ΣiJijSi z (tk-1)+gj+sgn(sj z (tk-1))·9pina}·tk / τ or Bj z (tk)={ΣiJiJSj z (tk-1)+gj+9pinb.Sj Z (tk-1)}·tk / τ is calculated, and the above variables Sj z The range of (tk) is -1≦sj z Define the function f so that (tk)≦1, and then Sj z (tk) = f(Bj z (tk), tk), and as the time step progresses from t=t0 to t=tm, the above variables Sj z approaching -1 or 1, and finally sj z If <0, then Sj zd =-1, Sj z If >0, then Sj zd Define the solution as =1.

[0068] The coefficient gpinb is, for example, 50% to 200% of the average value of |Jij|. Also, for the local field gj in the problem setting, a correction term δgj' can be added to gj' only for a certain site j' to increase the magnitude of gj' only for that site j'. Also, the correction term δgj' is, for example, 10% to 100% of the average value of |Jij|.

[0069] Next, the basic principles of the annealing machine realized by the transportation plan creation support device 100 will be described by starting from a quantum mechanical description and moving to a classical form.

[0070] The problem of searching for the ground state of the Ising spin Hamiltonian given by equation (3) includes a classification of problems called NP-hard problems and is known to be a useful problem (Reference: F. Barahona, "On the computational comp lexity of Isingspin glass models," J. Phys. A: Math. Gen. 15, 3241 (1982).).

[0071] [Formula 3] JPEG0007591007000003.jpg27170Jij and gj are the task setting parameters, and σ^ Z is the z component of the Pauli spin matrix and takes eigenvalues ​​of ±1. i,j represent the spin site. Ising spin is a variable that can only take the value ±1, and in equation (3), σ^ z Since the eigenvalues ​​of are ±1, it is an Ising spin system.

[0072] The Ising spin in equation (3) does not have to be a literal spin; it can be anything physically possible as long as the Hamiltonian is described by equation (3).

[0073] For example, it is possible to correspond the use or non-use of each delivery route by each vehicle (described later) to ±1, or to correspond the high and low of a logic circuit to ±1, or to correspond the vertical and horizontal polarization of light to ±1, or to correspond the phase of 0,π to ±1.

[0074] In the method exemplified here, similar to adiabatic quantum computing, a computation system is prepared in the ground state of the Hamiltonian given by equation (4) at time t=0.

[0075] [Formula 4] JPEG0007591007000004.jpg28170γ is a proportionality constant determined by the magnitude of the external field uniformly applied to all sites j, and σ^j x is the x-component of the Pauli spin matrix. If the calculation system is the spin itself, the external field means the magnetic field.

[0076] Equation (4) corresponds to the application of a transverse magnetic field, and the ground state is when all spins point in the x direction (γ>0). The Hamiltonian in the problem formulation was defined as an Ising spin system with only the z component, but equation (4) includes the x component of spin. Therefore, the spin in the calculation process is not Ising but vectorial (Bloch vector). At t=0, we start with the Hamiltonian in equation (4), but gradually change the Hamiltonian as time t progresses, and finally obtain the ground state as the Hamiltonian described by equation (3) is obtained as the solution.

[0077] [Formula 5] JPEG0007591007000005.jpg25170Here, σ^ represents the three components of the Pauli spin matrix as a vector. The ground state is when the spin is oriented in the magnetic field direction, and can be written as <σ^>=B / |B|, where <·> is the quantum mechanical expectation value. In an adiabatic process, the ground state is always maintained, so the spin direction always follows the direction of the magnetic field.

[0078] The above discussion can be extended to multi-spin systems. At t=0, the Hamiltonian is given by equation (4). This applies a uniform magnetic field Bj X = γ is applied. At t>0, the x component of the magnetic field gradually weakens and Bj X = γ(1-t / τ). For the z component, the effective magnetic field is given by equation (6) due to the spin-spin interaction.

[0079] [Formula 6] JPEG0007591007000006.jpg33170The spin direction is <σ^ z > / <σ^ X >, so if the spin direction follows the effective magnetic field, the spin direction is determined by equation (7).

[0080] [Formula 7] JPEG0007591007000007.jpg27170Equation (7) is a quantum mechanical description, but it takes the expectation value, so unlike equations (1) to (6), it is a relationship related to classical quantities.

[0081] In a classical system, there is no nonlocal correlation (quantum thread) of quantum mechanics, so the spin direction should be completely determined by the local field at each site, and formula (7) determines the behavior of the classical spin system. In a quantum system, formula (7) is transformed due to the existence of nonlocal correlation, but this will be described later. Here, we will describe the classical system determined by formula (7) to explain the basic form of the invention.

[0082] Figure 9 shows a timing chart for obtaining the ground state of the spin system. Since the description in this figure is of classical quantities, the spin of site j is expressed by sj rather than σ^j. Accordingly, the effective magnetic field Bj in this figure is a classical quantity. At t=0, a rightward effective magnetic field Bj is applied to all sites, and all spins Sj are initialized to the right.

[0083] As time t passes, the magnetic field in the z-axis direction and the spin-spin interaction are gradually added, and eventually the spin becomes +z or -z, and the z component of the spin Sj becomes sj z =+1 or is -1. Ideally, time t should be continuous, but in the actual calculation process, it can be made discrete to improve convenience. The discrete case is described below.

[0084] The spin shown here is a vectorial spin because it includes not only the z component but also the x component. Its vectorial behavior can also be understood from Figure 9. The y component has not appeared so far, but this is because the external field direction is taken to be the xz plane, so there is no y component of the external field, and therefore <σ^ Y This is because >=0.

[0085] The spin of the computational system is assumed to be a three-dimensional vector of magnitude 1 (called a Bloch vector, which can describe the state as a point on a sphere), but the axis selection in the example shown in Figure 9 requires that only two dimensions be considered (the state can be described as a point on a circle).

[0086] Also, since γ is constant, Bj x (t)>0(γ>0) or Bj x (t)<0(γ<0). In this case, the two-dimensional spin vector can be described only by a semicircle, and Sj z If you specify Sj z Thus, in this example, spin is a two-dimensional vector, but it can also be written as a one-dimensional continuous variable with a range of [-1,1].

[0087] In the timing chart of FIG. 9, the effective magnetic field is calculated for each site at time t=tk, and the spin direction at t=tk is calculated using this value according to equation (8).

[0088] [Formula 8] JPEG0007591007000008.jpg22170Equation (8) is equation (7) rewritten in terms of classical quantities, so there is no <·> symbol. Next, the effective magnetic field at t=tk+l is calculated using the value of the spin at t=tk. The effective magnetic field at each time can be written specifically as equations (9) and (10).

[0089] [Formula 9] JPEG0007591007000009.jpg35170[Formula 10] JPEG0007591007000010.jpg33170Then, the spin and effective magnetic field are alternately determined according to the procedure shown diagrammatically in the timing chart of FIG.

[0090] In the classical system, the magnitude of the spin vector is 1. In this case, each component of the spin vector is tanθ=Bj z (tk) / Bj x Using the parameter θ defined by (tk), Sj z (tk) = sinθ, Sj x It is written as (tk)=COSθ.

[0091] If we rewrite this, Sj z (tk) = sin(arctan(Bj z (tk) / Bj x (tk))), Sj x (tk) = cos(arctan(Bj z (tk) / Bj x (tk))).

[0092] As is clear from equation (9), Bj x The only variable in (tk) is tk, and τ and γ are constants. Therefore, Sj z (tk) = sin(arctan(Bj z (tk) / Bj x (tk))) and Sjx(tk)=cos(arctan(Bj z (tk) / Bj x (tk))) is Bj z Sj as a function of (tk) and tk z (tk) = f1(Bj z (tk), tk) and Sj x (tk)=f2( Bj z A generalized expression such as (tk),tk) is also possible.

[0093] Since spin is described as a two-dimensional vector, Sj z (tk) and sj x (tk) are two components, but Bj z If (tk) is determined based on equation (10), then Sjx (tK) is not necessary.

[0094] This is Sj z This corresponds to the fact that the spin state can be described by (tk) alone. zd , Sj zd =-1 or 1, Sj z If (τ)>0, then Sj zd =1, Sj z If (τ)<0, then Sj zd =-1.

[0095] FIG. 10 shows a flowchart summarizing the solution algorithm for the Ising model described above. Here, tm=τ. Each step s1 to s9 in the flowchart corresponds to a process at a certain time in the timing chart of FIG. 9, which spans from time t=0 to t=τ. That is, steps s2, s4, and s6 in the flowchart correspond to the above equations (9) and (10) at t=t1, tk+l, and tm, respectively. The final solution is obtained in step s8 by calculating sj z <0 then Sj zd =-1, Sj z If >0, then Sj zd = 1 (s9).

[0096] So far, we have shown how the Ising model 400, which is a specific problem, is solved when it is expressed by equation (3). Next, we will explain how the Ising model 400, which is a specific problem, is expressed by equation (3) including the local field gj and the variable interaction Jij (i, j = 1, 2, ..., N) by giving a concrete example.

[0097] The specific problem here, that is, the Ising model 400, for example, for an objective function including, as terms, the evaluation values ​​of the candidate delivery routes for each evaluation criterion and a function for constraint conditions that is minimized when the second constraint condition is satisfied, the spin is whether or not each vehicle uses each delivery route, and the sensitivity between the variables in the function for constraint conditions is set as the strength of the interaction between the spins.

[0098] In this case, the local field gj is assumed to be set as the influence on the objective function of the values ​​of the variables in, for example, the evaluation values ​​of the candidate delivery route for each evaluation criterion and the constraint function that is minimized when the second constraint is satisfied.

[0099] Through the above considerations, the variable interactions Jij and the local fields gj (between each term of the above-mentioned objective function) are specifically set, and the delivery route for each vehicle is identified through a search for the ground state of the Ising model 400 represented by equation (3), that is, a search for the ground state in which the above-mentioned objective function is minimized.

[0100] The only calculation performed by the Ising model and the annealing method is to "minimize the objective function." Therefore, if there are constraints that must be satisfied when minimizing the objective function, they must be added to the objective function in some way.

[0101] for example, [Formula 11] Consider the constraint JPEG0007591007000011.jpg24170. If we convert this constraint into a function that is minimized when the constraint is satisfied, we get the following formula:

[0102] [Formula 12] JPEG0007591007000012.jpg24170The squared part is always a positive value, so this formula will only reach a minimum when the squared part is 0. The only time the content is 0 is when ΣXi-A=0, so if you solve the optimization problem that minimizes this function, you will automatically obtain a solution that satisfies ΣXi=A.

[0103] Furthermore, for example, in the above-mentioned annealing method, the items to be used as constraint conditions must also be included in the objective function, so that the objective function and the constraint conditions are treated with the same importance.

[0104] For example, consider the following optimization problem:

[0105] [Formula 13] JPEG0007591007000013.jpg49170If we change this to a formulation for annealing, we get the following:

[0106] [Formula 14] JPEG0007591007000014.jpg25170Here, P and Q are constants and are factors that determine which term is to be minimized with priority. For example, if you want to minimize all three terms equally (i.e., solve the problem without biasing the strength of the constraints), you would set values ​​to balance the terms, such as making P equal to Q.

[0107] On the other hand, if the problem is posed such that "the constraint in the third term is not as important as the constraint in the third term," then the desired solution can be obtained by making the value of P, which is the coefficient of the term that is important, greater than the value of Q.

[0108] In this way, the annealing method makes it possible to assign priorities to constraint conditions and set less important constraint conditions to be "satisfied as much as possible."

[0109] Various settings related to the Ising model are appropriately performed in accordance with the conditions and information of this embodiment, in a form that follows existing general concepts.

[0110] The above-described functions realized by the transportation plan creation support device 100 are realized by the calculation unit 104 reading and executing each program stored in the memory 103 or the storage unit 101. Each program can be recorded on a recording medium and distributed, for example. Note that the transportation plan creation support device 100 may be realized, in whole or in part, by using virtual information processing resources provided by using virtualization technology, process space separation technology, or the like, such as a virtual server provided by a cloud system. Also, all or in part of the functions provided by the transportation plan creation support device 100 may be realized by a service provided by a cloud system via an API (Application Programming Interface), for example. Next, the processing performed by the transportation plan creation support system 1 will be described.

[0111] <Transportation planning support processing> 11 is a flow diagram illustrating an example of a transportation plan creation support process performed by the transportation plan creation support system 1. The transportation plan creation support process is executed, for example, when the transportation plan creation support device 100 receives a predetermined start instruction from the user terminal 200 via a user input.

[0112] First, the transportation plan creation support device 100 sets information on conditions related to the transportation work (s10). For example, the transportation plan creation support device 100 receives information corresponding to the vehicle information table 310, the base information table 320, the order information table 330, the travel time information table 340, the cost information table 350, and the evaluation criterion information table 360 ​​from the user terminal 200 via input from the user, and stores the received information in each corresponding table.

[0113] Next, the transportation plan creation support device 100 executes a delivery route candidate creation process s20, which uses mathematical programming to calculate suitable combinations of delivery routes (hereinafter referred to as delivery route candidates) that satisfy the first constraint condition that each vehicle is required to comply with under each evaluation criterion, based on the information stored in s10. The delivery route candidate creation process s20 will be described in detail later.

[0114] Then, the transportation plan creation support device 100 executes an optimal delivery route creation process s30 that creates an optimal combination of delivery routes for all vehicles (hereinafter referred to as optimal delivery route) based on each delivery route candidate created in the delivery route candidate creation process s20, using the Ising model 400. Details of the optimal delivery route creation process s30 will be described later.

[0115] The transportation plan creation support device 100 displays (s40) the information on the optimum delivery route created in the optimum delivery route creation process s30 on the transportation plan creation support device 100 or the user terminal 200. This ends the transportation plan creation support process. Each of the above processes will now be described in detail.

[0116] <Delivery route candidate creation process> The delivery route candidate creation process can be performed by a conventional method of solving a mathematical programming problem (optimization problem) using mathematical programming. For example, a delivery route can be created by the method shown in Patent Document 1. Hereinafter, a method of creating a delivery route will be briefly described based on Patent Document 1.

[0117] FIG. 12 is a flow diagram showing an example of a process for determining a delivery route candidate that satisfies the first constraint condition and optimizes the delivery cost when the evaluation criterion is the delivery cost.

[0118] As shown in the figure, the delivery route candidate creation unit 111 judges whether or not a delivery route for a vehicle under consideration, created using a genetic algorithm, satisfies the weight or volume of the cargo or the type of vehicle, etc., in the first constraint condition (S201). For example, the delivery route candidate creation unit 111 refers to the maximum capacity loading capacity 314 and the maximum weight loading capacity 315 in the vehicle information table 310, and the enterable vehicles 325 in the base information table 320.

[0119] If the delivery route satisfies the first constraint condition (S201: YES), the delivery route candidate creation unit 111 executes the processing of S202, and if the delivery route does not satisfy any of the first constraint conditions (S201: NO), the delivery route candidate creation unit 111 terminates the processing (for example, considers other delivery routes for the vehicle under consideration).

[0120] In S202, the delivery route candidate creation unit 111 searches for the best delivery route. Specifically, the delivery route candidate creation unit 111 calculates the evaluation value of the delivery cost for each delivery route candidate at each insertion point of the delivery destination on the delivery route based on the evaluation criterion information table 360. Then, the delivery route candidate creation unit 111 creates a new delivery route by inserting the delivery destination (order) to be newly assigned at the position where the evaluation value of the delivery cost (cost based on time and cost based on distance) is maximum. Then, the delivery route candidate creation unit 111 selects the best route as a tentative delivery route candidate.

[0121] The delivery route candidate creation unit 111 judges whether the selected tentative delivery route candidate (the delivery time schedule indicated by the selected tentative delivery route candidate) satisfies the time constraint in the first constraint condition (S203). For example, the delivery route candidate creation unit 111 refers to the maximum operating time 312 in the vehicle information table 310, the available collection time slot 335 and the available unloading time slot 336 in the order information table 330, and the operating time slot 322 in the base information table 320.

[0122] If the tentative delivery route candidate satisfies the time constraints (S203: YES), the delivery route candidate creation unit 111 executes the processing of S204, and if the tentative delivery route candidate does not satisfy the time constraints (S203: NO), the delivery route candidate creation unit 111 terminates the processing (for example, considers other delivery routes for the vehicle under consideration).

[0123] In S204, the delivery route candidate creation unit 111 determines whether the tentative delivery route candidate satisfies the distance constraints. For example, the delivery route candidate creation unit 111 refers to the longest operating distance 313 in the vehicle information table 310.

[0124] If the tentative delivery route candidate satisfies the distance constraint (S204: YES), the delivery route candidate creation unit 111 stores the tentative delivery route candidate as a formal delivery route candidate (S205) and ends the process. If the tentative delivery route candidate does not satisfy the distance constraint (S204: NO), the delivery route candidate creation unit 111 ends the process (e.g., consider another delivery route for the vehicle under consideration).

[0125] The delivery route candidate creation unit 111 performs the above-mentioned processing for each of the multiple evaluation criteria in the evaluation criteria information table 360 ​​for each vehicle under consideration.

[0126] (Delivery route options) 13 is a diagram showing an example of a delivery route candidate created by the delivery route candidate creation process s20. The delivery route candidates are optimized in terms of evaluation value for each evaluation criterion, and include a delivery route candidate 1301 for each vehicle having a high evaluation value for the evaluation criterion related to the vehicle's load ("load"), a delivery route candidate 1302 for each vehicle having a high evaluation value for the evaluation criterion related to the vehicle's operating time ("operating time"), and a delivery route candidate 1303 for each vehicle having a high evaluation value for the evaluation criterion related to the vehicle's travel distance ("travel distance").

[0127] For example, candidate delivery routes with high evaluation values ​​for the evaluation criterion related to vehicle load capacity include a route in which "Vehicle 1" transports luggage via "Route 1-1", a route in which "Vehicle 2" delivers luggage via "Route 2-1", a route in which "Vehicle 3" delivers luggage via "Route 3-1", and a route in which "Vehicle n" delivers luggage via "Route n-1".

[0128] In addition, candidate delivery routes with high evaluation values ​​for the evaluation criterion related to vehicle operating time include a route in which "Vehicle 1" transports luggage via "Route 1-2", a route in which "Vehicle 2" delivers luggage via "Route 2-2", and a route in which "Vehicle n" delivers luggage via "Route n-2".

[0129] In addition, delivery route candidates with high evaluation values ​​for the evaluation criterion related to the vehicle's travel distance include a route in which "Vehicle 1" transports luggage via "Route 1-3" and a route in which "Vehicle 2" delivers luggage via "Route 2-3".

[0130] The above-described delivery route candidate creation process s20 is just an example. The delivery route candidate creation process s20 is not limited to the method described here, and may be any process that solves an optimization problem using an objective function having a term of a first constraint condition in a mathematical programming method such as linear programming (for example, determines the value of each variable when the value of the objective function is the minimum value or the minimum value) to sequentially calculate the delivery route and delivery time schedule of each vehicle based on each evaluation criterion.

[0131] However, because mathematical programming creates delivery routes for each vehicle in sequence, there may be a bias in the evaluation value depending on the vehicle (for example, the evaluation value of a vehicle calculated in the latter stage of the process tends to be low). However, this problem can be improved by the optimal delivery route creation process s30, as described below. Next, the optimum delivery route creation process s30 will be described.

[0132] <Optimal delivery route creation process> 14 is a flow diagram illustrating details of the optimum delivery route creation process s30. The optimum delivery route creation unit 112 first acquires the second constraint condition, weight value, and target value of the evaluation criterion used in the Ising model 400 (s301).

[0133] For example, the optimal delivery route creation unit 112 receives the total number of vehicles P used for delivery via user input from the user terminal 200. As will be described later, when P=0, it is not a constraint condition but a parameter for minimizing the total number of vehicles.

[0134] In addition, the optimum delivery route creation unit 112 receives each weight value (vehicle number weight value A, evaluation value weight value B, route number weight value C, evaluation value deviation weight value D) described later from the user terminal 200 via user input.

[0135] In addition, the optimal delivery route creation unit 112 sets a target value A of the evaluation value related to the evaluation index k as a second constraint condition. k is received from the user terminal 200 via user input.

[0136] Furthermore, the optimal delivery route creation unit 112 acquires the route and evaluation value of the vehicle to be used in the Ising model 400 (s303).

[0137] For example, the optimal delivery route creation unit 112 acquires the combination of each route for each vehicle (candidate delivery route) calculated in the candidate delivery route creation process s20, the delivery time schedule for the candidate delivery route, and the evaluation value of each evaluation criterion related to the candidate delivery route.

[0138] 15 is a diagram showing an example of data acquired by the optimal delivery route creation unit 112. As shown in the figure, the optimal delivery route creation unit 112 acquires a combination 1501 of each route of each vehicle, and an evaluation value 1502 relating to the operating time, an evaluation value 1503 relating to the travel distance, and an evaluation value 1504 relating to the number of orders handled, which correspond to the combination of routes of the vehicle. Note that the combination 1501 of each route of each vehicle is managed as a binary value 1505 indicating whether or not to adopt it in the Ising model 400, as will be described later.

[0139] Next, as shown in s303 of FIG. 14, the optimal delivery route creation unit 112 creates the following equation (15) as an Ising model 400 based on the second constraint condition, weight value, target value, vehicle route, and evaluation value acquired in s301 and s303.

[0140] [Formula 15] JPEG0007591007000015.jpg18170

[0141] where x i indicates whether route i is used (specifically, whether it is used (1) or not (0)). The binary-valued term indicates that the spin ik is the evaluation value of evaluation index k for route i, R j point is a set of routes that pass through destination j, A is a weighting value (vehicle number weighting value A) that indicates the degree of importance of bringing the total number of vehicles used for delivery closer to the target value P, B is a weighting value (evaluation value weighting value B) that indicates the degree of importance of optimizing the evaluation value, C is a weighting value (route number weighting value C) that indicates the degree of importance of having only one vehicle pass through each base, and D is a weighting value (route number weighting value C) that indicates the degree of importance of having only one vehicle pass through each base. k is the target value A of the evaluation value of the route (vehicle) related to the evaluation index k k is a weighting value (evaluation value deviation weighting value D) that indicates the degree of importance placed on reducing the deviation from

[0142] The first term on the right side of equation (15) indicates the deviation of the number of vehicles used for delivery from the target value P. When the Hamiltonian H is in the ground state, the value of this term is minimized (i.e., the number of vehicles used for delivery approaches the target value P). Note that when P=0, the above formula minimizes the total number of vehicles. In other words, this term becomes the term to be optimized when P is not 0, and becomes the second constraint term (constraint function term) when P=.

[0143] The second term on the right-hand side of equation (15) is the sum of the evaluation values ​​of each path according to each evaluation criterion. The value of this term is maximized when the Hamiltonian H is in the ground state.

[0144] The third term on the right-hand side of equation (15) is the second constraint term (constraint function term) that indicates the number of vehicles that pass through each base (that collects cargo at each base) minus 1. When the Hamiltonian H is in the ground state, the value of this term is minimized (i.e., it approaches the state where only one vehicle collects cargo at each base).

[0145] The fourth term on the right side of equation (15) is the target value A of the evaluation value related to the evaluation criterion k of the route (vehicle). k is the second constraint term (constraint function term) that indicates the degree of deviation from . When the Hamiltonian H is in the ground state, the value of this term is minimized (i.e., each evaluation value approaches the target value E).

[0146] In the delivery route candidate creation process s20, routes are created sequentially for each vehicle, so there is a bias in the evaluation value (here, operating time) between routes, for example, route A is 5 hours, route B is 1 hour, etc. Therefore, by incorporating the fourth term on the right side above into the Ising model 400, it is possible to eliminate such bias between routes.

[0147] In addition, the target value A for a certain evaluation criterion l l When is set to 0, the Ising model 400 does not consider bias in the evaluation values ​​related to the evaluation criterion l between the vehicles. This makes it possible to, for example, ensure that there is no bias in the work hours, but allow for bias in the number of processed orders.

[0148] In this way, the Ising model 400 performs calculations with the second constraint as a penalty. Therefore, the solution in the Ising model 400 is a solution in a direction that satisfies the second constraint, but does not necessarily satisfy the second constraint. In other words, the second constraint in the Ising model 400 is not a strict condition like the first constraint in the delivery route candidate creation process s20, but a loose constraint with a less strict level of compliance (deviation may be allowed).

[0149] (Optimal delivery route) Fig. 16 is a diagram showing an example of an optimal delivery route created by the optimal delivery route creation process s30. This optimal delivery route is the result of identifying the route of each vehicle when the evaluation values ​​of all the evaluation criteria of all the vehicles are optimal overall from the delivery route candidates related to the four evaluation criteria shown in Fig. 13 (optimal delivery route).

[0150] For example, "Vehicle 1" is identified as using route 1601, "Route 1-2," which has a high evaluation value for operating time; "Vehicle 3" is identified as using route 1603, "Route 3-1," which has a high evaluation value for load capacity; and "Vehicle n" is identified as using route 1604, "Route n-2," which has a high evaluation value for operating time.

[0151] 14 and 16 on the screen of the transportation plan creation support device 100 or the user terminal 200. The result output unit 113 may also display the delivery schedule created in the delivery route candidate creation process s20, which is associated with the optimal delivery route.

[0152] As described above, the transportation plan creation support device 100 of the present embodiment creates a combination of delivery routes (delivery route candidates) for each of a plurality of evaluation criteria, which optimizes the evaluation value based on a predetermined evaluation criterion while satisfying the first constraint condition, which is a constraint condition for delivery and does not allow deviation (delivery route candidate creation process s20 using mathematical programming), and calculates an Ising model 400 in which the use or non-use of each delivery route for each vehicle is set as a spin, and the sensitivity between variables in the constraint condition function is set as the strength of interaction between the spins, for an objective function including, as terms, the evaluation value of the delivery route candidate for each evaluation criterion and a constraint condition function that is minimized when the second constraint condition, which is a constraint condition related to the evaluation value and allows deviation, is satisfied (optimum delivery route candidate creation process s30), and outputs information on the use or non-use of each route for each vehicle to an output device based on the results of the calculation.

[0153] That is, the transportation plan creation support device 100 of the present embodiment strictly adheres to the transportation plan. An optimal delivery route candidate for each delivery that complies with the first constraint condition to be met is created for multiple evaluation criteria, and then, using the Ising model 400 that performs calculations while regarding the second constraint condition as a penalty, a delivery route that optimizes the evaluation value of each evaluation criterion as a whole while satisfactorily satisfying the second constraint condition is specified, thereby creating an optimal delivery route. Furthermore, in this case, the transportation plan creation support device 100 determines the delivery route based on multiple evaluation values ​​using the Ising model 400 instead of optimization using mathematical programming or the like, thereby making it possible to create a well-balanced delivery route that has a high evaluation value based on multiple evaluation criteria as a whole without imposing unnecessarily strict constraints on each evaluation value.

[0154] In this way, the transportation plan creation support device 100 of this embodiment can create information on a transportation plan that takes into consideration both constraint conditions that must be strictly observed and constraint conditions that are desirable to observe but do not necessarily require strict observation, while using the Ising model 400. That is, the transportation plan creation support device 100 of this embodiment can support rapid creation of a transportation plan having many constraints while observing those constraints according to their contents.

[0155] For example, when formulating a transportation plan, there are various strict constraints, such as vehicle loading limits, time-specific cargo delivery, or the types of vehicles that can be used at each base station. Even in such cases, appropriate transportation plan information can be quickly created.

[0156] In addition, the transportation plan creation support device 100 of this embodiment creates delivery route candidates that optimize an evaluation value based on the evaluation criteria while satisfying a first constraint condition related to the type of vehicle, restrictions on the amount of cargo, or the time or distance required for delivery, and calculates an Ising model 400 related to an objective function that includes, as terms, the evaluation value of the delivery route candidate for each evaluation criterion and a constraint condition function that is minimized when the deviation of the evaluation value from a predetermined target value is minimized.

[0157] In this way, the transportation plan creation support device 100 can optimize transportation operations by setting vehicle type, cargo amount restrictions, or delivery time or delivery distance as first constraint conditions that must be strictly observed, and setting the evaluation value to be a specified target value as a second constraint condition that does not necessarily require strict observation.

[0158] In addition, the transportation plan creation support device 100 of the present embodiment calculates an Ising model 400 including, as a term, a constraint condition function that is minimized when the deviation of the number of vehicles in the delivery route candidate from the target value P is minimized.

[0159] In this way, the transportation plan creation support device 100 sets the type of vehicle and the number of vehicles used for delivery to be the target value P as a second constraint that does not necessarily require strict compliance, thereby enabling flexible and rational operation of transportation operations.

[0160] In addition, the transportation plan creation support device 100 of the present embodiment calculates an Ising model 400 that includes the number of vehicles in the delivery route candidate as a term.

[0161] As a result, the transportation plan creation support device 100 can optimize (specifically, minimize) the number of vehicles in the delivery route candidates by using the Ising model 400, thereby enabling efficient management of transportation operations.

[0162] In addition, the transportation plan creation support device 100 of this embodiment uses a weight value A set according to the evaluation criterion k for each evaluation criterion. k The Ising model 400 is calculated, which includes the evaluation value of the delivery route candidate to which the value is assigned as a term.

[0163] In this way, by using the weight values, the Ising model 400 can calculate the weights for each evaluation criterion. In particular, the Ising model 400 can perform tuning using such weight values ​​at a much higher speed than, for example, a mathematical programming method, even when there are many evaluation criteria, so that detailed policies for the transportation operations can be set quickly.

[0164] The present invention is not limited to the above-described embodiment, and can be implemented using any components without departing from the scope of the present invention. The above-described embodiments and modifications are merely examples, and the present invention is not limited to these contents as long as the features of the invention are not impaired. In addition, although various embodiments and modifications have been described above, the present invention is not limited to these contents. Other aspects conceivable within the scope of the technical idea of ​​the present invention are also included in the scope of the present invention. [Explanation of symbols]

[0165] 100 Transportation plan creation support device, 111 Delivery route candidate creation unit, 112 Optimal delivery route creation unit, 113 Result output unit

Claims

1. A device for supporting the creation of a transportation plan for delivering each package by a plurality of vehicles, comprising: a storage unit that stores a first constraint condition that is a constraint condition in the delivery and from which deviation is not permitted, and a second constraint condition that is a constraint condition related to each evaluation value based on each of a plurality of evaluation criteria in the delivery and from which deviation is permitted; and a delivery route candidate creation process for creating a combination of delivery routes for each of the vehicles as delivery route candidates for each of the multiple evaluation criteria, the combination satisfying the first constraint condition and optimizing an evaluation value based on the evaluation criteria; An optimal delivery route creation process that calculates an Ising model in which the use of each delivery route of each vehicle is set as a spin, and the sensitivity between variables in the constraint condition function is set as the strength of interaction between the spins, for an objective function including an evaluation value of the delivery route candidate for each of the evaluation criteria and a constraint condition function that is minimized when the second constraint condition is satisfied; and a result output process of outputting information indicating whether each vehicle has used each route to an output device based on a result of the calculation. Transportation planning support device.

2. the storage unit stores, as the first constraint condition, a constraint condition related to a type of the vehicle, a limit on the load of the luggage by the vehicle, or a time or distance required for the delivery, and stores, as the second constraint condition, a constraint condition related to a target value of the evaluation value; The calculation unit is In the delivery route candidate creation process, a combination of delivery routes for each of the vehicles is created as a delivery route candidate for each of the multiple evaluation criteria, such that an evaluation value based on the evaluation criteria is optimized while satisfying constraints related to the type of the vehicle, restrictions on the load of the cargo by the vehicle, or the time or distance required for the delivery; In the optimal delivery route creation process, the Ising model is calculated for an objective function including an evaluation value of the delivery route candidate for each evaluation criterion and a constraint condition function that is minimized when the deviation of the evaluation value from the target value is minimized. The transportation plan creation support device according to claim 1.

3. The storage unit stores, as the second constraint condition, a constraint condition related to a target value of the number of vehicles in the delivery; In the optimal delivery route creation process, the calculation unit further calculates the Ising model for an objective function including, as a term, a constraint condition function that is minimized when the deviation of the number of vehicles in the delivery route candidate from the target value is minimized. The transportation plan creation support device according to claim 1.

4. The calculation unit further calculates the Ising model for an objective function including a number of vehicles in the delivery route candidate as a term in the optimal delivery route creation process. The transportation plan creation support device according to claim 1.

5. The calculation unit, in the optimal delivery route creation process, calculates the Ising model for an objective function including, as a term, an evaluation value of the delivery route candidate to which a weight set according to each evaluation criterion is assigned for each evaluation criterion. The transportation plan creation support device according to claim 1.

6. A method for supporting creation of a transportation plan for a plurality of vehicles to deliver packages, comprising the steps of: An information processing device, A first constraint condition that is a constraint condition in the delivery and from which deviation is not permitted; a second constraint condition for each evaluation value based on each of the plurality of evaluation criteria in the first constraint condition, from which deviation is permitted; a delivery route candidate creation process for creating a combination of delivery routes for each of the vehicles as delivery route candidates for each of the multiple evaluation criteria, the combination satisfying the first constraint condition and optimizing an evaluation value based on the evaluation criteria; An optimal delivery route creation process that calculates an Ising model in which the use of each delivery route of each vehicle is set as a spin, and the sensitivity between variables in the constraint condition function is set as the strength of interaction between the spins, for an objective function including an evaluation value of the delivery route candidate for each of the evaluation criteria and a constraint condition function that is minimized when the second constraint condition is satisfied; and outputting a result of the calculation to an output device information indicating whether each vehicle has used each route based on the result of the calculation. A method for supporting the creation of transportation plans.

7. The information processing device, storing, as the first constraint condition, a constraint condition related to the type of the vehicle, a limit on the load of the cargo carried by the vehicle, or a time or distance required for the delivery, and storing, as the second constraint condition, a constraint condition related to a target value of the evaluation value; In the delivery route candidate creation process, a combination of delivery routes for each of the vehicles is created as a delivery route candidate for each of the multiple evaluation criteria, such that an evaluation value based on the evaluation criteria is optimized while satisfying constraints related to the type of the vehicle, restrictions on the load of the cargo by the vehicle, or the time or distance required for the delivery; In the optimal delivery route creation process, the Ising model is calculated for an objective function including an evaluation value of the delivery route candidate for each evaluation criterion and a constraint condition function that is minimized when the deviation of the evaluation value from the target value is minimized. The transportation plan creation support method according to claim 6.

8. The information processing device, As the second constraint condition, a constraint condition regarding a target value of the number of vehicles in the delivery is stored; In the optimal delivery route creation process, the Ising model is calculated for an objective function including, as a term, a constraint function that is minimized when the deviation of the number of vehicles in the delivery route candidate from the target value is minimized. The transportation plan creation support method according to claim 6.

9. The information processing device, In the optimal delivery route creation process, the Ising model is further calculated for an objective function including the number of vehicles in the delivery route candidate as a term. The transportation plan creation support method according to claim 6.

10. The information processing device, In the optimal delivery route creation process, the Ising model is calculated for an objective function including, as a term, an evaluation value of the delivery route candidate to which a weight set according to each evaluation criterion is assigned for each evaluation criterion. The transportation plan creation support method according to claim 6.

Citation Information

Patent Citations

  • Method and device for searching optimal solution of allocation and despatching scheduling program problem

    JP1996153085A

  • Pattern collating device

    JP1998055439A

  • Transportation planning device, transportation planning system, and transportation planning method

    JP2020181282A

  • Delivery plan creation method, operation method, and delivery plan creation device

    JP2021111276A

  • Information processing device, information processing method and information processing program

    JP2021144511A