JP2001014296A - Car allocation and delivery planning method, computer- readable recording medium recording car allocation and delivery planning program, and car allocation and delivery planning device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make effectively visitable each visiting place by means of a stochastic car allocation/delivery plan. SOLUTION: A required time distribution calculation part 1 is prepared together with a too-early/late penalty calculation part 2, a total cost calculation part 3 which calculates the total cost necessary for visiting all customers, a visiting sequence deciding part 4 which sets visiting sequence of customers so as to minimize the total cost under a prescribed condition and an arithmetic control part 5 which controls each of these said parts. In such a constitution, the too-early/late penalties are calculated from a required time distribution, i.e., a probability distribution of increase/decrease of the required time and then the trucks are allocated and a visiting sequence of customers is set in consideration of the calculation results of the said penalties. As a result, the influence of an unexpected situation such as a traffic accident can be minimized.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle delivery planning method, a computer readable recording medium storing a vehicle delivery planning program, and a vehicle delivery planning device.
In particular, the present invention relates to a vehicle delivery planning method, a vehicle delivery planning device, and a computer-readable recording medium that stores a vehicle delivery planning program for determining a vehicle allocation and a visit order when making a round trip to a plurality of destinations. The present invention can be applied to, for example, a case where cargo is delivered to a plurality of destinations, and conversely, a case where cargo is collected from a plurality of destinations.

[0002]

2. Description of the Related Art With the development of the economy, a sophisticated logistics system has been required in order to meet the diversified and sophisticated needs of consumers. In particular, in collection and delivery in cities, time designation is becoming stricter, and so-called just-in-time transportation is becoming widespread. In addition, it is considered that the perception of the time required for road users to move will increase more and more in the future due to the development of information and communication technology. Also, in the logistics in cities, reduction of logistics costs and labor saving are issues for the industry, and various problems such as traffic congestion, exhaust gas, noise, and energy saving are encountered for the traffic environment.

[0003] Under such circumstances, in recent years, there has been a movement in the logistics industry to construct an advanced vehicle dispatching / planning system in order to rationalize vehicle dispatching / delivery, and several studies have been made. For example, the present inventor has previously proposed a model that integrates an optimal dispatch plan of a pickup truck in a city and a dynamic traffic simulation. In this model, when a pickup truck goes around multiple customers to pick up or deliver cargo, the fixed and time costs required for truck operation and the penalty cost for arriving or arriving late at a specified time are considered. This is to minimize the sum. Here, a so-called deterministic vehicle delivery plan was used, assuming that the time required for movement between two points is always constant.

[0004]

However, in the above-described model, it is assumed that the time required for movement between two points is always constant. May not be able to collect and deliver cargo efficiently.

[0005] The present invention has been made in view of such a point, and the time required to travel between visit places or between customers due to various factors such as traffic accidents, traffic regulations, congestion, and breakdowns. It is an object of the present invention to provide a vehicle delivery planning method, a vehicle readable recording medium recording a vehicle delivery planning program, and a vehicle delivery planning device capable of efficiently visiting each visiting place and customer even if there is a change. And

Further, the present invention stochastically processes a transportation route and the like by considering a required time distribution for traveling, as compared with a conventional deterministic vehicle allocation and delivery planning system.
The purpose is to construct a so-called stochastic vehicle delivery system. Another object of the present invention is to reduce costs required for vehicle delivery by incorporating stochastic cost minimization processing and the like in order to minimize various costs for transportation.

[0007]

In the stochastic vehicle dispatching plan according to the present invention, an optimal vehicle dispatching plan can be formulated in consideration of the uncertainty of the required time. In other words, when the time required is considered as one fixed value and the vehicle dispatch and delivery plan is planned (deterministic dispatch and delivery plan), when the required time increases due to a sudden traffic accident or traffic congestion, Delays delivery of cargo. On the other hand, if the probability of an increase in the required time due to a traffic accident or traffic congestion is considered in advance in a probabilistic manner, the delay in the required time can be reduced even in such a situation. For example, in the case where a package is delivered to almost the same customer every day, such a stochastically optimal dispatch and delivery plan makes the total cost more economic than planning deterministically.

In order to solve the above-mentioned problems, according to a first solution of the present invention, a required time distribution calculation for calculating a required time distribution required for a moving means to move from one visited place to the next visited place. A total cost calculating step of calculating a total cost required to visit a plurality of visiting locations based on the required time distribution calculated by the required time distribution calculating step; and the total cost under predetermined conditions. A vehicle dispatching step including a step of deciding a visit order for assigning moving means to visit each visiting place and setting an order of visiting places to be visited by each moving means so that the total cost computed by the computing step is minimized; Provide a delivery planning method.

According to the present invention, the total cost is calculated in consideration of the required time distribution indicating the variation of the required time, so that even if an unexpected situation such as a traffic accident occurs, it is hardly affected by the unexpected cost. Further, in the present invention, the total cost is calculated in consideration of the time difference between the time of actually arriving at the visiting place and the estimated time of arrival, thereby setting the order of visiting such that the user can arrive at a time close to the expected time of arrival. be able to.
Further, in the present invention, the time difference between the time of arrival at the visited place and the predetermined expected arrival time is multiplied by a unit early arrival penalty cost or a unit late penalty cost, and the value is multiplied by a predetermined time distribution to perform time integration. By doing so, the penalty cost for early arrival and late arrival is calculated, so that the penalty cost can be calculated by a relatively simple calculation method.

Further, in the present invention, the required time distribution can be calculated with high accuracy by obtaining the average value and the standard deviation of the required time distribution based on the past history data.
Further, in the present invention, the required time distribution can be calculated by a simple calculation method without any past history data. In the present invention, the accuracy of the probabilistic model is improved by calculating the total cost taking into account the fixed cost, the traveling time cost, the cargo handling time cost, and the like. further,
In the present invention, a practical result can be obtained by determining the visit order under the condition that the load capacity of the vehicle does not exceed the limit amount.

According to a second solution of the present invention, a required time distribution calculating step for calculating a required time distribution required for moving means to move from one visited place to the next visited place, and the required time distribution calculating step Based on the required time distribution calculated according to the total cost calculation step of calculating the total cost required to visit a plurality of visit locations, the total cost calculated by the total cost calculation step under predetermined conditions is Computer readable recording of a vehicle dispatching and planning program comprising, to a minimum, the assignment of transportation means to visit each visit location and a visit order determination step to set the order of visit locations visited by each transportation means A simple recording medium is provided.

According to a third solution of the present invention, a required time distribution calculating section for calculating a required time distribution required for moving means to move from one visited place to the next visited place, and the required time distribution calculating section Based on the required time distribution calculated by the total cost calculation unit that calculates the total cost required to visit a plurality of visiting places, the total cost calculated by the total cost calculation unit under predetermined conditions is To minimize
There is provided a vehicle dispatching / planning apparatus that includes a visit order determining unit that sets an assignment of traveling means to visit each visiting place and an order of visiting places visited by each traveling means.

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, a vehicle dispatch planning method, a computer-readable recording medium recording a vehicle dispatch planning program, and a vehicle dispatch planning apparatus according to the present invention will be specifically described with reference to the drawings.

First, as an example, a description will be given of an example in which a plurality of customers are visited in order by a moving means such as a truck, and cargo is collected from each customer. Each truck departs from the place of departure (hereinafter referred to as a depot), visits customers in a specified order, collects the cargo, and finally returns to the depot and collects the cargo several times. . The truck is
It is assumed that only one car is assigned to one customer and all the cargo of the customer is loaded in one collection.

FIG. 1 is an explanatory diagram showing an example of a truck action pattern. The triangles in the center of FIG. 1 indicate a truck terminal or a warehouse (hereinafter, referred to as a depot). FIG. 1 shows, as an example, a case where ten customers are assigned to two trucks, each truck sequentially visits the customers to collect and deliver cargo, and then returns to the depot.

Specifically, the truck T1 is a depot → customer
N8 → N2 → N9 → N4 → N3 → N10 → Depot, truck T
2 goes around in the order of depot → customer N1 → N5 → N7 → N6 → depot. Such a tour is only an example, and there are many orders for the tour.

However, in practice, there is a limit to the number of trucks and the loading capacity, and it is economically advantageous to travel as many customers as possible with one truck, since the traveling distance becomes shorter. There are many. Therefore, it is necessary to obtain an optimal solution of the order of visiting trucks in consideration of these various conditions.

Also, in the example of FIG. 1, it is assumed that there are roads only in the east, west and north and south directions, and it is not taken into account that a truck travels in an oblique direction. It has become. Even in such a case, the present invention can be applied by calculating the traveling distance according to the actual road shape.

Further, in this embodiment, if the truck arrives at the destination earlier than the designated time, the truck waits until the designated time, and an early arrival penalty is imposed in proportion to the time difference from the estimated arrival time. Conversely, in the case of being late, a late penalty is imposed in proportion to the time difference from the estimated arrival time.

It is assumed that customers are scattered at a plurality of locations, for example, 10 to 20, and customer information (eg, location, cargo volume, cargo handling time, designated arrival time, early arrival and late arrival unit penalty costs). Is known. The number of trucks that can be used is, for example, up to ten, and information on each transportation means (for example, loading capacity, fixed cost, unit traveling time cost, unit handling time cost) is known.

FIG. 2 is a block diagram showing a schematic configuration of the vehicle dispatching planning apparatus according to the present invention. This vehicle allocation and delivery planning device includes a required time distribution calculation unit 1, an early arrival / late arrival penalty calculation unit 2, a total cost calculation unit 3, a visit order determination unit 4,
And an arithmetic control unit 5.

The required time distribution calculating section 1 calculates a variation in required time required for moving means such as a truck from one point to the next point (hereinafter referred to as required time distribution). The required time distribution calculation unit 1 determines an expected value and a standard deviation, for example, assuming that past history data relating to the required time is a normal distribution, and calculates the required time distribution based on the normal distribution. The required time distribution calculation unit 1 can calculate the required time based on data received from a road condition management system such as an intelligent transportation system, as described later.

The early arrival / late arrival penalty calculating section 2 calculates an early arrival / late arrival penalty when the truck arrives early or late. Early arrival / late arrival penalty cost calculator 2
The time difference between the time of arrival at the visited place and the predetermined expected arrival time, the unit early arrival penalty cost per unit time when arriving at the visited place early, and the unit late penalty cost per unit time when arriving late at the visited place And penalty costs for early arrival and late arrival are calculated based on the arrival time distribution to each visited location. The early arrival / late arrival penalty cost calculation unit 2 calculates the arrival time distribution to each visited location based on the predetermined time distribution calculated by the predetermined time distribution calculation unit. For example, the early arrival / late arrival penalty cost calculation unit 2 multiplies the product of the time difference between the arrival time at the visited place and the expected arrival time and the unit early arrival / late arrival penalty cost by the arrival time distribution to perform time integration. To calculate penalty costs for early arrival and late arrival.

The total cost calculation unit 3 includes a required time distribution calculation unit 1
Based on the required time distribution calculated according to the above, the total cost required to visit and visit a plurality of or all customers and visiting places is calculated. The total cost calculation unit 3 sets the total cost as a result of adding at least one of the fixed cost, the traveling time cost, and the cargo handling time cost, and the penalty cost for early arrival or late arrival.
Further, the total cost calculator 3 calculates the total cost including the penalty cost calculated by the early arrival / late arrival penalty cost calculator 2.

The visit order determining unit 4 assigns a moving means for visiting each visiting place and determines whether each moving means is necessary so that the total cost calculated by the total cost calculating unit 3 is minimized under predetermined conditions. Set the order of visiting places to visit. The visit order determination unit 4 can set a predetermined condition that the loaded amount of the vehicle does not exceed the limit amount. In addition, the arithmetic control unit 5
The control of each section described above is performed.

FIG. 3 is a flowchart showing the processing operation of the vehicle dispatching and planning method according to the present invention. Hereinafter, the operation of this embodiment will be described based on this flowchart.

First, calculation conditions are set (step S).
1). The calculation conditions include customer information and transportation means information (for example, road network, location of depot and customer, customer cargo demand, time designation, number of trucks, truck loading capacity), and the like. If necessary,
These conditions are set in any combination.

Next, the required time distribution calculator 1 calculates the required time distribution among the customers (step S2). Specifically, the required time distribution is assumed to be a normal distribution, and the expected value μ of the required time is calculated based on the distance between the customers and the traveling speed of the truck. The average traveling speed of the truck is constant during one cycle.

The expected value μ is expressed by equation (1). Here, di, j is the distance (km) between the visited places i, j, and S is the running speed (km / h) of the truck. μ = di, j / S (1)

It is assumed that the relationship between the standard deviation σ of the required time distribution and the expected value μ is expressed by the following equation (2). σ = α · μ (2) In the equation (2), α is a parameter (%) representing the magnitude of the standard deviation. In the present embodiment, in order to simplify the calculation, for example, α was changed in the order of 0.05, 0.1, 0.15, and 0.2, and the standard deviation σ was determined.

When there is past history data relating to the required time, the expected value μ and the standard deviation σ can be directly obtained using these history data. Alternatively, if there is no history data, the following (3)
The standard deviation σ may be calculated based on an empirical formula shown in the formula. It should be noted that the constants (0.36, 0.821) in the formulas are merely examples, and the present invention is not limited thereto. σ (standard deviation) = 0.36 × μ ^0.821 (3)

When the calculation of the required time distribution is completed, a sequence X =
_{{X 1, x 2, ···} , x k, ···, x m} initializing a (step S3). Here, m is the number of available trucks, and x _k constituting the sequence X is a sequence in which the customers visited by the truck k are arranged in the order of visit. For example, in the case of the example of FIG. 1, since two trucks are used,
The sequence X = {x ₁ , x ₂ }, and x ₁ = (0,8,2,9,4,3,10,
0), x ₂ = (0,1,5,7,6,0). Note that “0” represents a depot, and the customers N1 to N10 in FIG. 1 are represented by numerical values of 1 to 10, respectively.

When the processing in step S3 is completed, the early arrival / late arrival penalty calculator 2 calculates the early arrival / late arrival penalty (step S4). The early arrival / late arrival penalty calculator 2 calculates the early arrival / late arrival penalty P for the customer i of the truck k based on the following equation (4).
Calculate _i (x _k ) (circle).

[0034]

(Equation 1)

[0035] ^{_{P i a (x k, t}} ) in equation (4) shows the arrival time distribution in the customer i track k. here,
A required time distribution expressed in expectation μ and standard deviation σ described above, arrival time distribution ^{_{P i a (x k, t}} ) and associated will be described. The required time distribution is the time required to move between the depot and the customer or between customers, and the arrival time distribution is the time required to move from the depot to a certain customer. For example,
In FIG. 1, the required time distribution between the depot and the customer N8 is the expected value μ.
It indicated by ₁ and a standard deviation sigma _1, the required time distribution between the customer N8 and customer N2 was shown in expectation mu ₂ and a standard deviation sigma _2. At this time, the arrival time distribution of the customer N2 is, for example, the sum of these required time distributions.

DU _i (x _k , t) indicates a penalty function at the customer i, and is represented by the following equation (5).

[0037]

(Equation 2)

In the equation (5), x _k is a sequence representing the assignment and visiting order of the track k, ts, i is the start time (minute) of the designated arrival time zone at the customer i, and te, i is the customer i End time (minutes) of the designated arrival time zone, Ce, i is a unit early arrival penalty cost (yen / min), and Cd, i is a unit late penalty cost (yen / min).

FIG. 4 is an explanatory diagram schematically showing the procedure for calculating the early arrival and late penalties. FIG.
(A) shows coefficients Ce, i and Cd, i in equation (5),
FIG. 4B is an arrival time distribution showing a required time distribution, and FIG.
(C) shows the calculation result of equation (5). FIG. 4 (C)
As shown in (5), by performing the calculation of equation (5), two curves, a curve a for early arrival penalty and a curve b for late penalty, are obtained.

When the processing in step S4 in FIG. 3 is completed,
Next, the total cost calculator 3 calculates the total cost C (X) (yen) required to visit all customers (step S5). In practice, for example, an optimal vehicle dispatch plan for a company is one that minimizes the actual total cost (eg, the sum of fixed costs, travel time costs, handling time costs, early arrival and late penalties, etc.). Conceivable. Thus, when the objective function is formulated as the expected total cost in consideration of the required time distribution, the following equation (6) is obtained. The total cost calculation unit 3
The total cost C (X) is calculated based on the equation (6).

[0041]

(Equation 3)

Here, C _{f, k} is a fixed cost (yen / vehicle) per truck k, and δ _k (x _k ) is a coefficient indicating whether or not the truck has been operated. 1
If it has not been activated, it will be 0. C _{t, k} is the unit traveling time cost of the truck k (yen / min. / Unit), C _{s, k} is the truck k
, T _k (x _k ) is the total running time of the truck k (minutes), and S _k (x _k ) is the total handling time of the truck k (minutes).

That is, the total cost is the cost according to the number of trucks in operation, the cost according to the running time of the truck,
It is expressed as the sum of the cost according to the total cargo handling time and the above-mentioned early arrival and late penalty costs.

Next, it is determined whether or not an appropriate total cost C (X) has been obtained (step S6). As the end determination, for example, the value of the total cost C (X) is compared with the value of the total cost C (X) of the previous time and the current time. The determination can be made based on whether the value is equal to or less than a predetermined threshold value. Here, if an appropriate total cost C (X) is not obtained,
That is, if it is determined that the processing has not been completed, the process returns to step S3 and the processing is repeated. On the other hand, when it is determined that the processing is to be ended, the process proceeds to the next step S7.

Next, within the range satisfying the condition of the following equation (7), the arithmetic control unit 5 sets the equation (4) so that the total cost C (X) shown in the equation (6) is minimized. By solving the vehicle assignment and distribution planning problem of the formulas (7) to (7), a sequence X representing the assignment of trucks and the order of visits by customers is determined (step S7). That is, the total cost calculation unit 3 satisfies the condition of the expression (7),
The step S is performed so that the total cost shown in the equation (6) is minimized.
The sequence X initially set in step 3 is changed. W _k (x _k ) ≦ Wmax, k (7) In the equation (7), W _k (x _k ) indicates the loading capacity (kg) of the truck k, and Wmax, k is the maximum loading capacity of the truck k. (kg).

In general, in such a vehicle dispatching and distribution plan, when the number of customers to be patroled is large, and when a visit time is specified, it is very complicated to strictly find an optimal solution. Therefore, it is preferable to use an approximate solution method to calculate an optimal solution. For example, one of the heuristic methods, Simulated Annealing (K
okubugata, H., Itoyama, H. et al., Vehicle routing meth
ods for city logistics operations, IFAC / IFIP / IFORS
Symposium on Transportation Systems, Chania, Gree
ce, eds.M. Papageorgiou & A. Pouliezos, pp. 755-76
0, 1997, etc.). In addition, as a method of obtaining an approximate solution, a genetic algorithm (Thangiah, S. R et al., A genet
ic algorithm system for vehicle routing with time
windows, Seventh IEEE International Conference on
Artificial Intelligence Applications, IEEE Compute
r Society Press, Los Alamitos, CA, pp.322-328, 199
1), Tabu Search (Potvin, J.
-Y., Kervahut, T. et al., The vehicle routing problemwi
th time windows; Part I: tabu search, INFORMS Jour
nal on Computing, 8, pp. 158-164, 1996.).

Next, evaluation of the stochastic delivery allocation plan model will be described. Here, as an example, a cost reduction rate (VSS, Value of Stochastic Solution) was used as an evaluation index of the stochastic model obtained by the above processing procedure. The VSS, resulting a cost C _F obtained by deterministic model using the average value of the required time as a constant value, as obtained by the procedure described above, by a stochastic model considering the standard deviation was the difference between the cost _{C P} (C
Is a value showing a ratio of _F -C _P).

FIG. 5 is an explanatory diagram of the cost reduction rate by the stochastic vehicle allocation distribution plan with respect to the average traveling speed. This average running speed is an average running speed based on the surrounding traffic and traffic congestion conditions, and is constant in this example. The area model was, for example, a square having a vertical and horizontal direction of 20 km each. In the regional model, the depot is located at the center and several 10-20 customers are randomly placed. As a delivery means, ten 4-ton trucks were used. In cases (1) to (4), the number of customers is 20, 18, 14, and 12, respectively. The test results for Cases (1) to (4) are white circles, black circles,
The graph is shown by a black triangle plot, and the average is shown by a black square plot.

According to the average graphs of cases (1) to (4), when the average traveling speed was reduced from 30 km / h to 18 km / h or less, the cost reduction rate showed a peak of 19%. Therefore, if the probabilistic vehicle allocation and delivery plan of the present embodiment is adopted, if the traffic conditions are deteriorated due to traffic congestion, accidents, construction, and the like, and the average traveling speed is reduced to, for example, 30 km / h or less, the cost reduction rate is reduced. It can be seen that it is reduced by a maximum of 19%. Therefore, the vehicle dispatch plan based on the stochastic model of the present embodiment can play an important role for selecting an efficient customer visit order.

As described above, in the present embodiment, taking into account that the time required for a truck to visit a customer fluctuates due to traffic conditions and the like, the required time distribution which is a probability distribution of the increase and decrease of the required time is taken into account. Calculates the early arrival and late penalties based on the above, and assigns trucks and sets the order of customer visits taking into account the calculation results, so that even if an accident such as a traffic accident occurs, Will not receive much. Further, when there are customers who visit frequently, the total cost required for a long-term (for example, one year) visit can be significantly reduced by optimizing the order of visits according to the processing in FIG. .

Furthermore, recently, the intelligent transportation system (I
With the spread of TS (Intelligent Transport Systems), detailed historical data on the required time distribution of trucks and passenger cars has become easily available. By analyzing such historical data with a computer,
The required time distribution in step S2 in FIG. 3 can be calculated in a short time and with high accuracy, and a highly reliable stochastic model can be obtained.

The ITS is a system mainly composed of information exchange performed between vehicles and roads and between vehicles using mobile communication and peripheral technologies. Applications of ITS include, for example, advanced navigation systems, automatic toll systems, support for safe driving, optimization and efficiency of traffic management and road management, and support for public transportation and pedestrians. In the present embodiment,
Utilizing such ITS, it is possible to obtain information about the travel history of a moving means such as a truck and to record it if necessary. For example, a histogram of the required time can be obtained by using ITS and converted into a normal distribution and input. In the past year, one week,
It is possible to easily, accurately and quickly obtain data on required time distribution in units of one day. Further, such a required time distribution can be used as an input of a probabilistic vehicle allocation delivery plan. In addition, every day using ITS,
By obtaining data at appropriate time intervals such as weekly or monthly, the required time distribution can be updated to the latest information, and a dynamic simulation can be performed. That is, the stochastic model obtained in the processing of FIG. 3 and the dynamic traffic simulation model may be combined to set the customer visit order.

In the above-described embodiment, an example has been described in which a customer is visited to collect a cargo. However, the present invention can be applied to the delivery of a cargo, and a customer is collected for a purpose other than the collection and delivery of a cargo. Applicable when visiting. Further, the means of transportation is not limited to a truck, and any suitable means including walking can be used. The present invention is also applicable, for example, to a case where a salesman visits a customer on a vehicle such as a passenger car or a motorcycle for business purposes.

In step S1 of FIG. 3, an example has been described in which the required time distribution is assumed to be a normal distribution. However, the distribution is not necessarily assumed to be a normal distribution, but may be assumed to be another curve distribution. Also, the unit in the above description is an example, and “
Instead of dollars, euros, hours instead of minutes,
Appropriate units such as tons and kg can be used in place of "table" such as week and day. Further, when calculating the total cost by using the equation (6), it is possible to add the time at which each track departs from the depot as a decision variable. In that case,
Equation (6) is modified as in equation (8).

[0055]

(Equation 4) Here, C (t ₀ , X): each departure time of each track is a vector t
Total cost in the case of ₀ (yen) t ₀ : Vector t ₀ = {t _1,0 , t _2,0 , t _3,0 ,..., Representing the time when the truck leaves the depot
_{t m, 0} t k,} 0: time track k departs the depot _{T k (t k, 0,} x k): departure time t _{k, 0} total running time of the track k, which left the depot at (min) P _i (t _{k, 0} , x _k ): Early arrival and late penalty for customer i of truck k that left the depot at departure time t _{k, 0} Other terms in equation (8) are as in equation (6) It is.

Here, for example, assuming that track 1 and track 2 are used, track 1 departs from the depot at 9:00,
If truck 2 leaves the depot at 10 o'clock, the vector t
₀ = {9, 10}. The time is in minutes, 10
It can be determined as appropriate, such as a minute unit or a 30 minute unit.

As described above, the system which minimizes the total cost by using the vector t ₀ representing the time of departure from the depot of each truck, the sequence X representing the allocation of each truck to the customer, and the order of visits as decision variables. Can be constructed. The vehicle allocation and delivery plan of the present invention as described above can be applied to such a system.

The above-mentioned vehicle delivery plan is realized by a vehicle delivery plan program, and the program can be provided by being recorded on a computer-readable recording medium.

[0059]

As described above in detail, according to the present invention, the total cost required to visit all the visited places is calculated based on the required time distribution, so that the probability of increase or decrease of the required time is determined in advance. The visit order can be set taking into account. Further, according to the present invention, even if the time required to travel between visited places or between customers varies due to various factors such as traffic accidents, traffic regulations, traffic jams, breakdowns, etc., the influence is not so much affected. Already, it is possible to efficiently visit each visiting place and customer.

Further, according to the present invention, a so-called stochastic theory in which a transportation route and the like are processed stochastically by taking into consideration a required time distribution for traveling as compared with a conventional deterministic dispatching and distribution planning system. A dynamic vehicle delivery system can be constructed. Further, according to the present invention, the cost required for vehicle delivery can be reduced by incorporating a stochastic cost minimization process or the like in order to minimize various costs for transportation.

Furthermore, according to the present invention, the required time distribution can be calculated easily and accurately with the recent spread and spread of ITS, so that it is necessary to move in consideration of an unexpected situation such as a traffic accident. Since the required time distribution is available, a stochastic vehicle dispatch plan can be implemented simply, accurately and quickly.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram showing an example of an action pattern of a truck.

FIG. 2 is a block diagram showing a schematic configuration of a vehicle allocation and delivery planning device according to the present invention.

FIG. 3 is a flowchart showing a processing operation of the vehicle allocation and delivery planning method according to the present invention.

FIG. 4 is an explanatory view schematically showing a procedure for calculating early arrival and late penalties.

FIG. 5 is an explanatory diagram of a cost reduction rate by a stochastic vehicle allocation delivery plan with respect to an average traveling speed.

[Explanation of symbols]

 1 Required time distribution calculation unit 2 Early arrival / late arrival penalty calculation unit 3 Visit order determination unit 4 Total cost calculation unit 5 Calculation control unit

Claims (12)

[Claims] 1. A required time distribution calculating step for calculating a required time distribution required for moving means to move from one visited place to the next visited place; and a required time distribution calculated by the required time distribution calculating step. A total cost calculating step of calculating a total cost required to visit a plurality of visiting places, and each visiting place such that the total cost calculated by the total cost calculating step under predetermined conditions is minimized. And a visit order determining step of setting an order of visiting places visited by each of the traveling means. 2. The total cost calculation step includes: at least one of a fixed cost, a traveling time cost, and a cargo handling time cost;
2. The vehicle allocation / delivery planning method according to claim 1, wherein an addition result of the early arrival or late arrival penalty cost is used as a total cost. 3. The time difference between the arrival time at the visiting place and the predetermined expected arrival time, the unit early arrival penalty cost per unit time when arriving at the visiting place early, and the unit time when arriving late at the visiting place. An early arrival / late arrival penalty cost calculating step of calculating an early arrival / late arrival penalty cost based on a unit late arrival penalty cost and an arrival time distribution to each visited location, wherein the total cost operation step is: The method according to claim 1 or 2, wherein the total cost is calculated including the penalty cost calculated in the early arrival / late arrival penalty cost calculation step. 4. The method according to claim 1, wherein the early arrival / late arrival penalty cost calculating step calculates an arrival time distribution to each visited location based on the predetermined time distribution calculated in the predetermined time distribution calculating step. 3. The vehicle delivery planning method according to 3. 5. The early arrival / late arrival penalty cost calculating step includes the step of multiplying a product of a time difference between an arrival time at a visiting place and an expected arrival time and a unit early arrival / late arrival penalty cost by an arrival time distribution. The vehicle allocation and delivery planning method according to claim 3 or 4, wherein the penalty costs for early arrival and late arrival are calculated by integration. 6. The required time distribution calculating step is characterized in that an expected value and a standard deviation are determined assuming that past history data relating to the required time is a normal distribution, and the required time distribution is calculated based on the normal distribution. The vehicle allocation / distribution planning method according to any one of claims 1 to 5, wherein 7. The required time distribution calculating step calculates the required time on the basis of data received from a road condition management system such as an intelligent transportation system. The vehicle allocation and delivery planning method described. 8. The vehicle dispatching and distribution method according to claim 1, wherein the visit order determining step is performed under a predetermined condition that a loaded amount of a vehicle does not exceed a limit amount. 9. The method according to claim 1, wherein said total cost calculation step further calculates a total cost according to each departure time of each transportation means. 10. The method according to claim 9, wherein said early arrival / late arrival penalty cost calculating step calculates an arrival time distribution according to each departure time of each transportation means. 11. A required time distribution calculating step for calculating a required time distribution required for moving means to travel from one visited place to the next visited place, and a required time distribution calculated by the required time distribution calculating step. A total cost calculating step of calculating a total cost required to visit a plurality of visiting places, and each visiting place such that the total cost calculated by the total cost calculating step under predetermined conditions is minimized. A computer-readable recording medium recording a vehicle dispatching and distribution planning program, comprising: a step of determining the order of visiting places to be visited by each of the traveling means; 12. A required time distribution calculating section for calculating a required time distribution required for moving means to move from one visited place to the next visited place, and a required time distribution calculated by the required time distribution calculating section. A total cost calculating unit for calculating a total cost required to visit a plurality of visiting places, and a visiting cost for each of the visiting places so that the total cost calculated by the total cost calculating unit under a predetermined condition is minimized. And a visit order determining unit for setting an order of visiting places to be visited by each of the traveling means, and an assignment of traveling means for visiting the traveling means.