CN103246969B - A kind of implementation method of logistics deployment and device - Google Patents

Abstract

The invention discloses a method and device for realizing logistics allocation. A mathematical model for realizing logistics allocation is generated by modeling logistics allocation constraint conditions, allocation station distance modeling, and logistics allocation algorithm modeling; on the basis of the mathematical model On top of that, the ant colony algorithm is used to allocate logistics. The technology for realizing logistics allocation in the present invention can generate a mathematical model for realizing logistics allocation through logistics allocation constraint modeling, allocation station distance modeling, and logistics allocation algorithm modeling; and apply ant colony on the basis of the mathematical model Algorithmic allocation of logistics, so it can generate a near-optimal logistics allocation plan in the smallest quantity and within the smallest range, which can greatly improve the efficiency of logistics allocation and save logistics allocation costs.

Description

A method and device for implementing logistics allocation

technical field

The invention relates to the Internet of Things technology, in particular to a method and device for realizing logistics allocation.

Background technique

The optimization of logistics allocation plan belongs to the category of combinatorial optimization, which is a non-deterministic polynomial (NP) problem, and the solution to this problem is a major research hotspot. In addition, because the actual state of logistics allocation is in dynamic changes, it is required that the combinatorial optimization algorithm should not only have good performance, but also be able to adapt to the changing characteristics of dynamic allocation. Some of the current logistics allocation algorithms use non-heuristic traditional algorithms, which have low efficiency in performance, while other heuristic algorithms with better performance do not take into account the changing characteristics of the allocation state, so they cannot be timely and dynamically The adjustment of the plan has led to the inefficiency of the logistics deployment plan.

Contents of the invention

In view of this, the main purpose of the present invention is to provide a method and device for realizing logistics allocation, which are used to realize high-efficiency logistics allocation and adapt to the dynamic characteristics of the logistics allocation process.

In order to achieve the above object, technical solution of the present invention is achieved in that way:

A method for implementing logistics allocation, the method comprising:

A mathematical model for realizing logistics allocation is generated by modeling logistics allocation constraint conditions, allocation station distance modeling, and logistics allocation algorithm; on the basis of the mathematical model, an ant colony algorithm is used to allocate logistics.

The method for modeling the logistics allocation constraints is:

Assume that the logistics allocation system has n logistics allocation stations in total, and each allocation station is denoted as a _i (i=1...n), at time T(t), the inventory of a certain logistics goods in the allocation station a _i is S _i (T (t)), and at this moment, the deployment station’s demand for the logistics goods is y _i (T(t)), which satisfies the state constraint condition S _i (T(t))≥y _i (T(t) ); (i=1...n).

The method also includes:

Logistics allocation management is carried out by establishing and maintaining a finite state machine for each logistics allocation station; the state machine function of logistics allocation station a _i is defined as: St _i (T(t)); where (i=1...n );

When the state machine cannot meet the state constraints at the corresponding time point in the future, the state transition of the state machine is performed by starting the logistics allocation plan to make it meet the state constraints.

The distribution station distance modeling includes absolute distance modeling and relative distance modeling; wherein,

In absolute distance modeling, set the distance between site a _i and site a _j as d _ij , and establish a two-dimensional matrix of distances between different sites:

When modeling relative distance, if there are no other stations on the path between station a _i and station a _j , then set the distance between station a _i and station a _j as d _ij ; if the distance between station a _i and station a _j There is a third station k in between, then d _ij =d _ik +d _kj ;

By analogy, the distance between all different sites is expressed as the sum of the indecomposable pairwise minimum distances between sites:

Among them, d′ _ij represents the relative distance between site a _i and site a _j , expressed as the relative value of the minimum distance between two sites.

The logistics allocation algorithm modeling is used to predict the future state, including the definition of the future state function and the definition of the logistics allocation matrix; wherein,

When defining the future state function, the future state prediction function is defined as: Num _i = fc(st _i (T(t)), y _i (T(t)), S _i (T(t)));

If Num _i > 0, it means that the allocation station a _i has remaining logistics goods, and the number is Num _i ; if Num _i = 0, it means that the allocation station a _i has no remaining logistics goods; when Num _i ≥ 0, it means that the allocation station a _i does not The logistics goods need to be supplemented through the logistics allocation program; if Num _i <0, it means that the allocation station a _i needs to start the logistics allocation program, and the number of logistics goods allocated to the allocation station a _i is the absolute value of Num _i ;

When defining the logistics allocation matrix, establish a two-dimensional logistics allocation matrix between the allocation stations. If the allocation station a _i does not need to allocate goods to the allocation station a _j , it is represented by 0; if it needs to be allocated, it is expressed as the value of the allocated logistics goods t _ij :

The method also includes:

use Indicates the sum of all logistics goods shipped to the deployment station a _j , which needs to meet the conditions:

That is, the logistics goods transported to the deployment station a _j must meet the requirements of the future state machine.

use Indicates the sum of the logistics goods transported from the deployment station a _i to all other deployment stations, and the conditions must be met: That is, the sum of the logistics goods transported by the deployment station a _i to all other deployment stations must meet the future state machine constraints of the deployment station a _i .

The process of applying the ant colony algorithm to allocate logistics on the basis of the mathematical model includes the following steps:

Step 1: Set the maximum number of cycles N _cmax , the number of ants m′, the initialization time slice t=0, the cycle control variable N _c =0, the ant cycle variable k=0, the tabu table tabu _k (k=1, 2, ..., m'), node dynamic adjustment table tabv _k (k=1, 2,..., m'), path table path _k (k=1, 2,..., m'), place m' ants in m On a starting node, make the pheromone τ _ij (t)=const on the connected arcs of different nodes, where const is a constant; put the nodes in all destination node sets into the node dynamic adjustment table tabv _k (k=1, 2 ,..., m'); the ant is a variable used for optimization in the ant colony algorithm;

Step 2: Number of cycles N _c =N _c+1 , if N _c >N _cmax , exit the cycle and jump to step 11;

Step 3: The number of ants k=k+1, if k>m, exit the k loop and jump to step 8;

Step 4: Determine the node where the ant is currently located, and detect whether the node belongs to the destination node set, if so, delete the destination node corresponding to the node from the node dynamic adjustment table tabv _k , otherwise jump to the next step;

Step 5: Find the path connected to the node where the ant is located, calculate the transition probability according to the state transition probability formula, and select the next node starting from this node;

Step 6: Modify the tabu table _tabuk , put the newly selected node into the tabu table; and update the pheromone;

Step 7: If the tabu table tabu _k already contains all the nodes in the set of starting nodes and the set of destination nodes, then the ant's current route selection is over, and jump to step 9, otherwise, jump to step 5;

Step 8: Calculate and record the sum of new pheromones added by the ants in this round of node selection. If the route selection of all ants in this round is over, go to step 9, otherwise go to step 3;

Step 9: Compare the sum of the newly added pheromones of the combined computing nodes selected by all ants in this round, take the combined sequence of computing nodes with the largest sum of newly added pheromones, and perform pheromone update on the pheromones on the corresponding path according to the pheromone update rules. update;

Step 10: Clear tabu table tabu _k and path table path _k , and jump to step 2;

Step 11: Take the deployment route represented by the maximum pheromone path as the optimal solution of the deployment scheme.

The method also includes:

Corresponding to the best logistics deployment plan that has been generated, apply real-time perception information to determine whether the plan needs to be dynamically adjusted. Delete the corresponding path in the dispatching path; for the dispatching task to be deleted, added, or modified currently, delete or add the corresponding node in the set of starting nodes and the set of destination nodes; determine the current node position of the dispatching vehicle or the upcoming arrival Node position, after the latest state update, generate the corresponding node state diagram; place all k ants on the current position node of the dispatching vehicle, and re-optimize the solution.

A device for implementing logistics allocation, the device includes a mathematical model generation module and a logistics allocation module; wherein,

The mathematical model generation module is used to generate a mathematical model for realizing logistics deployment through modeling of logistics deployment constraints, modeling of distances from deployment stations, and modeling of logistics deployment algorithms;

The logistics allocation module is used for applying the ant colony algorithm to allocate logistics on the basis of the mathematical model.

The mathematical model generation module includes a logistics allocation constraint modeling unit for:

Assume that the logistics allocation system has n logistics allocation stations in total, and each allocation station is denoted as a _i (i=1...n). At the future time T(t), the inventory of certain logistics goods in the allocation station a _i is S _i ( T(t)), and at this moment, the deployment station’s demand for the logistics goods is y _i (T(t)), which satisfies the state constraint condition S _i (T(t))≥y _i (T(t )); (i=1...n).

The logistics allocation constraint modeling unit is also used for:

Logistics allocation management is carried out by establishing and maintaining a finite state machine for each logistics allocation station; the state machine function of logistics allocation station a _i is defined as: St _i (T(t)); where (i=1...n );

When the state machine cannot satisfy the state constraint condition at the corresponding time point in the future, the state transition of the state machine is performed.

The mathematical model generation module includes a distribution station distance modeling unit for absolute distance modeling and relative distance modeling; wherein,

In absolute distance modeling, set the distance between site a _i and site a _j as d _ij , and establish a two-dimensional matrix of distances between different sites:

When modeling relative distance, if there are no other stations on the path between station a _i and station a _j , then set the distance between station a _i and station a _j as d _ij ; if the distance between station a _i and station a _j There is a third station k in between, then d _ij =d _ik +d _kj ;

By analogy, the distance between all different sites is expressed as the sum of the indecomposable pairwise minimum distances between sites:

Among them, d′ _ij represents the relative distance between site a _i and site a _j , expressed as the relative value of the minimum distance between two sites.

The mathematical model generation module includes a logistics allocation algorithm modeling unit for predicting the future state, including the definition of a future state function and the definition of a logistics allocation matrix; wherein,

When defining the future state function, the future state prediction function is defined as: Num _i = fc(st _i (T(t)), y _i (T(t)), S _i (T(t)));

If Num _i > 0, it means that the allocation station a _i has remaining logistics goods, and the number is Num _i ; if Num _i = 0, it means that the allocation station a _i has no remaining logistics goods; when Num _i ≥ 0, it means that the allocation station a _i does not The logistics goods need to be supplemented through the logistics allocation program; if Num _i <0, it means that the allocation station a _i needs to start the logistics allocation program, and the number of logistics goods allocated to the allocation station a _i is the absolute value of Num _i ;

When defining the logistics allocation matrix, establish a two-dimensional logistics allocation matrix between the allocation stations. If the allocation station a _i does not need to allocate goods to the allocation station a _j , it is represented by 0; if it needs to be allocated, it is expressed as the value of the allocated logistics goods t _ij :

The logistics allocation algorithm modeling unit is also used for:

use Indicates the sum of all logistics goods shipped to the deployment station a _j , which needs to meet the conditions:

That is, the logistics goods transported to the deployment station a _j must meet the requirements of the future state machine.

use Indicates the sum of the logistics goods transported from the deployment station a _i to all other deployment stations, and the conditions must be met: That is, the sum of the logistics goods transported by the deployment station a _i to all other deployment stations must meet the future state machine constraints of the deployment station a _i .

When the logistics allocation module applies the ant colony algorithm to allocate logistics on the basis of the mathematical model, it is used to perform the following steps:

Step 1: Set the maximum number of cycles N _cmax , the number of ants m′, the initialization time slice t=0, the cycle control variable N _c =0, the ant cycle variable k=0, the tabu table tabu _k (k=1, 2, ..., m'), node dynamic adjustment table tabv _k (k=1, 2,..., m'), path table path _k (k=1, 2,..., m'), place m' ants in m On a starting node, make the pheromone τ _ij (t)=const on the connected arcs of different nodes, where const is a constant; put the nodes in all destination node sets into the node dynamic adjustment table tabv _k (k=1, 2 ,..., m'); the ant is a variable used for optimization in the ant colony algorithm;

Step 2: Number of cycles N _c =N _c+1 , if N _c >N _cmax , exit the cycle and jump to step 11;

Step 3: The number of ants k=k+1, if k>m, exit the k loop and jump to step 8;

Step 4: Determine the node where the ant is currently located, and detect whether the node belongs to the destination node set, if so, delete the destination node corresponding to the node from the node dynamic adjustment table tabv _k , otherwise jump to the next step;

Step 5: Find the path connected to the node where the ant is located, calculate the transition probability according to the state transition probability formula, and select the next node starting from this node;

Step 6: Modify the tabu table _tabuk , put the newly selected node into the tabu table; and update the pheromone;

Step 7: If the tabu table tabu _k already contains all the nodes in the set of starting nodes and the set of destination nodes, then the ant's current route selection is over, and jump to step 9, otherwise, jump to step 5;

Step 8: Calculate and record the sum of new pheromones added by the ants in this round of node selection. If the route selection of all ants in this round is over, go to step 9, otherwise go to step 3;

Step 9: Compare the sum of the newly added pheromones of the combined computing nodes selected by all ants in this round, take the combined sequence of computing nodes with the largest sum of newly added pheromones, and perform pheromone update on the pheromones on the corresponding path according to the pheromone update rules. update;

Step 10: Clear tabu table tabu _k and path table path _k , and jump to step 2;

Step 11: Take the deployment route represented by the maximum pheromone path as the optimal solution of the deployment scheme.

The logistics allocation module is also used for:

Corresponding to the best logistics deployment plan that has been generated, apply real-time perception information to determine the execution degree of the plan. For the tasks that have been deployed, delete the corresponding nodes from the set of starting nodes and the set of destination nodes, and start from the optimal scheduling path. Delete the corresponding path in ; for the current dispatching task to be deleted, added, or modified, delete or add the corresponding node in the starting node set and destination node set; determine the current node position of the dispatching vehicle or the upcoming node position , after the latest state update, generate the corresponding node state diagram; place all k ants on the current position node of the dispatching vehicle, and re-optimize the solution.

The technology for realizing logistics allocation in the present invention can generate a mathematical model for realizing logistics allocation through logistics allocation constraint modeling, allocation station distance modeling, and logistics allocation algorithm modeling; and apply ant colony on the basis of the mathematical model Algorithmic allocation of logistics, so it can adapt to the dynamic characteristics of the logistics allocation process, and generate a near-optimal logistics allocation plan within the minimum quantity and range, which can greatly improve the efficiency of logistics allocation and achieve high-efficiency logistics allocation. And save logistics deployment costs.

Description of drawings

Fig. 1 is a schematic flow diagram of realizing logistics allocation according to an embodiment of the present invention;

Fig. 2 is a diagram of a device for implementing logistics allocation according to an embodiment of the present invention.

detailed description

The present invention adopts a dynamic and improved ant colony algorithm for dynamic optimization of deployment schemes, and the algorithm is a heuristic algorithm. Heuristic algorithm, also known as intelligent computing, also known as "soft computing", is an algorithm that people are inspired by the laws of nature and imitate and solve problems according to the laws of nature. This type of algorithm has good global optimization performance, strong robustness, and strong versatility , suitable for parallel processing, etc., and because of the advantages of efficient optimization performance and no need for problem-specific information, it has been widely used in computer science, optimal scheduling, transportation problems, combinatorial optimization and other fields. Ant colony algorithm (ant colony algorithm) was first proposed by Italian scholars Dorigo, Maniezzo and others in the early 1990s. Ant colony algorithm is another heuristic search algorithm applied to combinatorial optimization problems after simulated annealing algorithm, genetic algorithm, tabu search algorithm, artificial neural network algorithm and other meta-heuristic search algorithms. Due to the characteristics of the ant colony algorithm itself, it can better adapt to the dynamic problems in the optimization process.

The establishment of an intelligent management system based on the Internet of Things technology makes it possible to obtain real-time information such as the inventory of items in the logistics allocation station, item names, types, real-time demand for items, and the specific location of logistics allocation vehicles. Through these real-time acquisitions Analysis and prediction of the sensory data can generate a near-optimal logistics allocation plan, and can also realize real-time update and dynamic adjustment of the logistics allocation plan.

The invention establishes a mathematical model and improves it based on the ant colony algorithm, and provides a logistics allocation scheme with good optimization effect and high efficiency. The present invention studies the logistics allocation problem based on the Internet of Things technology. Firstly, the mathematical model for logistics deployment is designed through the modeling of logistics allocation constraints, the distance modeling of allocation stations, and the modeling of logistics allocation algorithms. Then, based on the mathematical model, the logistics allocation problem is studied, including the application of ant colony Algorithm research on static logistics allocation algorithm and dynamic logistics allocation algorithm.

Specifically, the ant colony algorithm is a heuristic optimization algorithm (also known as intelligent computing, also known as soft computing), which has the characteristics of efficient optimization performance, global optimization performance, strong robustness, and strong versatility. It is widely used in computer science, optimal scheduling, transportation problems, combinatorial optimization, etc.

It should be noted that logistics allocation is very important, because only the timely allocation of required goods can ensure normal operation and improve the quality of logistics allocation. The application of Internet of Things technology makes it possible to obtain real-time perception information and predict information. The present invention applies forecasting technology, real-time perception and other technologies to generate a near-optimal logistics allocation plan within the minimum quantity and minimum range, which can save logistics allocation costs to a large extent.

When performing mathematical modeling, the following operations can be performed:

Modeling of logistics allocation constraints:

Assume that the logistics allocation system has n logistics allocation stations in total, and each allocation station is denoted as a _i (i=1...n). At a certain time T(t) in the future, the inventory of certain logistics goods in the allocation station a _i is S _i (T(t)), and at this moment, the demand of the dispatching station for the logistics goods is y _i (T(t)), in order not to affect the smooth progress of production and transactions, the system must meet the following state constraints:

S _i (T(t))≥y _i (T(t)); (i=1...n)(1)

In order to better explain and solve the problem, the theory of finite state machine is introduced. Logistics allocation management is carried out by establishing and maintaining a finite state machine for each logistics allocation station. The information contained and maintained by the finite state machine includes: time, the logistics goods inventory of the corresponding logistics allocation station at a certain time point, and the logistics goods demand of the corresponding logistics allocation station at this time point. When the state machine cannot satisfy the state constraint condition (1) at the corresponding time point in the future, the state transition of the state machine is required. The conditional functions that can promote the state change of the state machine of the logistics deployment station include f, g, etc.; among them,

g: logistics allocation function;

f: other influencing factors.

The state machine function of logistics deployment station a _i is defined as: St _i (T(t)); where (i=1...n).

Distribution station distance modeling:

The modeling of the distance between logistics deployment sites can be divided into absolute distance modeling and relative distance modeling.

1) Absolute distance modeling

Set the distance between site a _i and site a _j as d _ij , and establish a two-dimensional matrix of distances between different sites. Due to the particularity of the road, the distance from site a _i to site a _j may be different from the distance from site a _j to site a _i , so the matrix formed by modeling is not a symmetric matrix:

2) Relative distance modeling

All logistics deployment sites form a large Internet network, some of which may be directly connected to each other, and some of which may also have other sites.

In order to express the relative distance between different sites more intuitively, a relative distance matrix is also established, and the following rules are set:

If there are no other stations on the path between station a _i and station a _j , then set the distance between station a _i and station a _j as d _ij , where d _ij can be an absolute distance value or a relative distance value.

If there is a third station k between station a _i and station a _j , then d _ij =d _ik +d _kj .

By analogy, the distance between all different sites is expressed as the sum of the indecomposable pairwise minimum distances between sites:

Among them, d′ _ij represents the relative distance between site a _i and site a _j , expressed as the relative value of the minimum distance between two sites.

Logistics allocation algorithm modeling:

To correctly and reasonably determine the logistics deployment plan, factors such as future inventory, demand, and status changes need to be considered. Therefore, the future state function needs to be established:

1) Future state function definition

The quantity of goods to be deployed at a certain point in the future can be predicted based on characteristics such as experience values, historical data, and future changes. The future state function is defined as: Num _i = fc(st _i (T(t)), y _i (T(t)), S _i (T(t)));

If Num _i > 0, it means that the deployment station a _i has remaining logistics goods, the number is Num _i ; if Num _i = 0, it means that the deployment station a _i has no remaining logistics goods. When Num _i ≥ 0, it means that the deployment station a _i does not need to supplement the logistics goods through the logistics deployment procedure; if Num _i <0, it means that the deployment station a _i needs to start the logistics deployment procedure, and the number of logistics goods deployed to the deployment station a _i is the absolute value of Num _i .

2) Definition of logistics allocation matrix

Establish a two-dimensional matrix of logistics allocation between allocation stations. If allocation station a _i does not need to allocate goods to allocation station a _j , it is represented by 0; if allocation is required, it is expressed as the value of the allocated logistics goods t _ij :

Can use Indicates the sum of all logistics goods shipped to the deployment station a _j , which needs to meet the conditions:

That is, the logistics goods transported to the deployment station a _j must meet the requirements of the future state machine.

Can use Indicates the sum of the logistics goods transported from the deployment station a _i to all other deployment stations, and the conditions must be met: That is, the sum of the logistics goods transported by the deployment station a _i to all other deployment stations must meet the future state machine constraints of the deployment station a _i .

When performing specific logistics allocation, static logistics allocation algorithms can be applied, and dynamic logistics allocation algorithms can also be applied. Specifically, when the target state machine of the logistics deployment station cannot meet the constraint conditions, it is necessary to start the logistics deployment system within the specified time for the deployment of goods. The static logistics allocation algorithm is an allocation algorithm formulated when the logistics allocation has not yet started; the dynamic logistics allocation algorithm is an adjustment algorithm for real-time modification of the original allocation plan when the allocation task is in progress, through the real-time perceived information to determine the situation changes.

The static logistics allocation algorithm generally includes start-up, solution of allocation scheme and scheme optimization:

1) Start-up of the static logistics allocation algorithm

According to the real-time perceived information, it is detected whether the corresponding state machine does not meet the constraint conditions at the allocation station in the future, and based on this, it is determined whether to start the static logistics allocation algorithm.

2) Deployment plan solution

The static logistics allocation algorithm solves the logistics allocation matrix under the constraint conditions, and at the same time needs to meet the minimum scope and the least transportation cost of the generated allocation plan, that is, the optimization problem of the allocation plan. The specific operation is: for each group of solutions that meet the conditions of the logistics deployment matrix that meets the constraints of the target state machine, the optimal calculation of the logistics deployment plan is performed on the group of solutions, so as to obtain the optimal deployment plan under the group of solutions. Finally, the optimal allocation schemes of all solutions are compared, and the global optimal allocation scheme is obtained.

Specifically, the solution of the deployment plan is to add the negative value calculated by the state prediction function to 0, record the transformation process at the same time, and generate a corresponding transformation plan. This transformation plan is a logistics deployment plan. The recording of the transformation process is realized through a two-dimensional matrix whose initialization state is all zero, and the corresponding addition operation only occurs between positive and negative values, and the vector is added once, and the corresponding addition is also in the corresponding allocation matrix. The position is marked, so that when a transformation is completed, the corresponding allocation matrix is generated. Different transformation processes will produce different allocation matrices, that is, different allocation schemes. As for which allocation scheme is the optimal scheduling scheme, it is necessary to further apply the allocation scheme optimization algorithm for further optimization, and finally generate a nearly optimal scheduling scheme.

3) Optimization of deployment plan

First generate an optimal transportation plan for each deployment scenario. For the sake of simplicity, it is assumed that the cost of logistics deployment is only related to the dispatching distance, and the same dispatching vehicle has the same cost for dispatching one piece of goods and deploying multiple pieces of goods at the same time. First of all, for the generated deployment plan matrix, apply the distance matrix and relative distance matrix to calculate whether there are repeated routes for the dispatch of goods between different stations. After merging, the allocation plan optimization problem is transformed into a special, limited condition Traveling Salesman Problem (TSP) after the repeated lines are merged.

TSP theoretically has an optimal solution, but no effective method has been found to obtain this value in practical applications. It can only gradually approach the optimal solution, that is, only the suboptimal solution closest to the theoretical optimal solution can be obtained. TSP The solution problem is a typical NP problem. The optimization problem of deployment plan is a special TSP with limited conditions, because the solution process of TSP has no restrictions on the order of node visits, while the optimization problem of deployment plan has restrictions on the order of visits, and it needs to start from the station where the goods are transferred. Arrive at the station where the goods are transferred.

The following uses the ant colony algorithm to solve the above special and limited TSP.

Let m' be the number of nodes in the starting node set, and the corresponding number of ants is also m', and m' ants are placed on m' starting nodes. Described ant is the variable that is used for optimization in the ant colony algorithm, and n ' represents the number of all nodes in the set of starting nodes and the set of destination nodes, and d _ij is the distance between node i and node j, and τ _ij (t) is The amount of pheromone on the path (i, j) at time t, the initial time τ _ij (t) takes a constant const, τ _ij (t) determines the calculation of the state transition probability. The following is the formula for calculating the state transition probability:

${\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}^{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\frac{{<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>}^{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}<span\; class="notranslate">\; \&times;</span>{<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; \&eta;</span>}}_{\mathrm{<span\; class="notranslate">\; ik</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>}^{\mathrm{<span\; class="notranslate">\; \&beta;</span>}}}{\underset{\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; \&Subset;</span>\mathrm{<span\; class="notranslate">\; allowed</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; is</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>}^{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}<span\; class="notranslate">\; \&times;</span>{<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; \&eta;</span>}}_{\mathrm{<span\; class="notranslate">\; is</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>}^{\mathrm{<span\; class="notranslate">\; \&beta;</span>}}}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span>{\mathrm{<span\; class="notranslate">\; allowed</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}\\ \mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; else</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

In formula (2), Indicates the state transition probability of the k-th ant transferring from node i to node j connected to i at time t, and allowed _k is the set of all nodes connected to node i. α is an information heuristic factor, which represents the relative importance of the trajectory, and reflects the role of the information accumulated by the ant during the movement of the ant. The larger the value, the more the ant is inclined to choose the path that other ants pass The path, the stronger the cooperation between ants. β is the expected heuristic factor, which represents the relative importance of visibility, and reflects the importance of heuristic information in the ant's choice of path during the movement process. η _ij (t) is a heuristic function, if the computing node has no historical information, then η _ij (t) is an empirical value.

At time t+n, the update rule of pheromone on the node (i, j) arc is as follows:

τ _ij (t+n)=(1-ρ)×τ _ij (t)+ρ×Δτ _ij ;

$\mathrm{<span\; class="notranslate">\; \&Delta;</span>}{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; m</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; \&Delta;\&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}^{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; ;</span>$

$\mathrm{<span\; class="notranslate">\; \&Delta;</span>}{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}^{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; =</span>\left\{\begin{array}{c}\frac{\mathrm{<span\; class="notranslate">\; Q</span>}}{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}}\\ \mathrm{<span\; class="notranslate">\; 0</span>}\end{array}\right.<span\; class="notranslate">\; ;</span>$

The kth ant passes through nodes i and j at time t and t+n, otherwise it is 0. Among them, τ _ij (t) represents the value of pheromone on the (i, j) arc at time t, Indicates the value of the pheromone amount on the computing node (i, j) newly felt by an ant at time t+n. ρ represents the pheromone update factor, 1-ρ represents the pheromone residual factor, here ρ represents the extent to which the computing node learns from the newly experienced τ _ij (t) value, and 1-ρ represents the extent to which the previous τ _ij is borrowed (t) value.

According to the one-to-one correspondence between the nodes in the starting node set and the destination node set, put the destination node corresponding to the starting node into the corresponding tabv _k (k=1, 2, . . . , m′).

In order to obtain the globally optimal allocation scheme, the following steps can be performed:

Step 1: Set the maximum number of cycles N _cmax , the number of ants m′, the initialization time slice t=0, the cycle control variable N _c =0, the ant cycle variable k=0, the tabu table tabu _k (k=1, 2, ..., m'), node dynamic adjustment table tabv _k (k=1, 2,..., m'), path table path _k (k=1, 2,..., m'), place m' ants in m ′ on a starting node, let the pheromone τ _ij (t) on the connected arcs of different nodes=const, where const is a constant, and the value of const should not affect the result of subsequent algorithm operation. Put into node dynamic adjustment table tabv _k (k=1,2,...,m');

Step 2: Number of cycles N _c =N _c+1 , if N _c >N _cmax , exit the cycle and jump to step 11;

Step 3: The number of ants k=k+1, if k>m, exit the k loop and jump to step 8;

Step 4: Determine the node where the ant is currently located, and detect whether the node belongs to the destination node set, if so, delete the destination node corresponding to the node from the node dynamic adjustment table tabv _k , otherwise jump to the next step;

Step 5: Find the path connected to the node where the ant is located, calculate the transition probability according to the state transition probability formula, and select the next node starting from this node, that is, select the calculation node determined by this node and the next node;

Step 6: Modify the tabu table _tabuk , put the newly selected node into the tabu table; and update the pheromone;

Step 7: If the tabu table tabu _k already contains all the nodes in the set of starting nodes and the set of destination nodes, then the ant's current route selection is over, and jump to step 9, otherwise, jump to step 5;

Step 8: Calculate and record the sum of the newly added pheromones of the nodes selected by the ants in this round. The sum of the newly added pheromones here is the sum of the newly added pheromones of each child node; Step 9, otherwise jump to step 3;

Step 9: Compare the sum of the newly added pheromones of the combined computing nodes selected by all ants in this round, take the combined sequence of computing nodes with the largest sum of newly added pheromones, and perform pheromone update on the pheromones on the corresponding path according to the pheromone update rules. update;

Step 10: Clear tabu table tabu _k and path table path _k , and jump to step 2;

Step 11: The deployment route represented by the maximum pheromone path is the optimal solution of the deployment scheme.

By solving the best transportation route and the lowest transportation cost for all scheduling schemes, the best scheduling scheme for this scheduling is finally obtained.

It can be seen from the above that the static logistics allocation algorithm is started before the logistics allocation program is started, and is used to generate the best allocation plan for this allocation. However, since the state of the system is always changing, even in the process of generating and implementing the current logistics allocation plan, it is necessary to make dynamic and real-time adjustments according to changes in information to generate a dynamic logistics allocation plan. The dynamic logistics allocation algorithm is based on the static logistics allocation algorithm, adding the dynamic design of the algorithm, dynamically deleting or adding nodes and changing the corresponding algorithm flow.

The dynamic logistics allocation algorithm can include the following operations:

Corresponding to the best logistics deployment plan that has been generated, apply real-time perception information to determine the execution degree of the plan. For the tasks that have been deployed, delete the corresponding nodes from the set of starting nodes and the set of destination nodes, and start from the optimal scheduling path. Delete the corresponding path in ; for the current dispatching task to be deleted, added, or modified, delete or add the corresponding node in the starting node set and destination node set; determine the current node position of the dispatching vehicle or the upcoming node position , after the latest status update, generate the corresponding node status graph;

In addition, during initialization, all k ants are placed on the current position node of the dispatching vehicle, and the aforementioned static logistics allocation algorithm is used to re-optimize the solution.

In combination with the above, it can be seen that the operation idea of the present invention to realize logistics allocation can represent the flow shown in Figure 1, which includes the following steps:

Step 110: Generate a mathematical model for logistics allocation by modeling logistics allocation constraint conditions, allocation station distance modeling, and logistics allocation algorithm modeling;

Step 120: Applying the ant colony algorithm to allocate logistics on the basis of the mathematical model.

In order to ensure that the above operations can be carried out smoothly, a device as shown in Figure 2 can be set up, which includes a mathematical model generation module and a logistics allocation module that can be connected; wherein, the mathematical model generation module can model through logistics allocation constraints, and the allocation station Distance modeling and logistics allocation algorithm modeling are used to generate a mathematical model for realizing logistics allocation; the logistics allocation module can apply ant colony algorithm to allocate logistics on the basis of the mathematical model.

It can be seen that the present invention can generate a mathematical model for logistics allocation by modeling logistics allocation constraint conditions, allocation station distances, and logistics allocation algorithms; on the basis of the mathematical model, an ant colony algorithm is used to allocate logistics.

The method for modeling the logistics allocation constraints can be:

Assume that the logistics allocation system has n logistics allocation stations in total, and each allocation station is denoted as a _i (i=1...n), at time T(t), the inventory of a certain logistics goods in the allocation station a _i is S _i (T (t)), and at this moment, the deployment station’s demand for the logistics goods is y _i (T(t)), which satisfies the state constraint condition S _i (T(t))≥y _i (T(t) ); (i=1...n).

It is also possible to manage logistics allocation by establishing and maintaining a finite state machine for each logistics allocation station; the state machine function of logistics allocation station a _i is defined as: St _i (T(t)); where (i=1 ...n);

When the state machine cannot meet the state constraints at the corresponding time point in the future, the state transition of the state machine is performed by starting the logistics allocation plan to make it meet the state constraints.

The delivery station distance modeling may include absolute distance modeling and relative distance modeling; wherein,

In absolute distance modeling, set the distance between site a _i and site a _j as d _ij , and establish a two-dimensional matrix of distances between different sites:

When modeling relative distance, if there are no other stations on the path between station a _i and station a _j , then set the distance between station a _i and station a _j as d _ij ; if the distance between station a _i and station a _j There is a third station k in between, then d _ij =d _ik +d _kj ;

By analogy, the distance between all different sites is expressed as the sum of the indecomposable pairwise minimum distances between sites:

Among them, d′ _ij represents the relative distance between site a _i and site a _j , expressed as the relative value of the minimum distance between two sites.

The logistics allocation algorithm modeling can be used to predict the future state, including the definition of the future state function and the definition of the logistics allocation matrix; wherein,

When defining the future state function, the future state prediction function is defined as: Num _i = fc(st _i (T(t)), y _i (T(t)), S _i (T(t)));

If Num _i > 0, it means that the allocation station a _i has remaining logistics goods, and the number is Num _i ; if Num _i = 0, it means that the allocation station a _i has no remaining logistics goods; when Num _i ≥ 0, it means that the allocation station a _i does not The logistics goods need to be supplemented through the logistics allocation program; if Num _i <0, it means that the allocation station a _i needs to start the logistics allocation program, and the number of logistics goods allocated to the allocation station a _i is the absolute value of Num _i ;

When defining the logistics allocation matrix, establish a two-dimensional logistics allocation matrix between the allocation stations. If the allocation station a _i does not need to allocate goods to the allocation station a _j , it is represented by 0; if it needs to be allocated, it is expressed as the value of the allocated logistics goods t _ij :

Alternatively, you can use Indicates the sum of all logistics goods shipped to the deployment station a _j , which needs to meet the conditions:

That is, the logistics goods transported to the deployment station a _j must meet the requirements of the future state machine.

use Indicates the sum of the logistics goods transported from the deployment station a _i to all other deployment stations, and the conditions must be met: That is, the sum of the logistics goods transported by the deployment station a _i to all other deployment stations must meet the future state machine constraints of the deployment station a _i .

The process of applying the ant colony algorithm to allocate logistics on the basis of the mathematical model may include the following steps:

Step 1: Set the maximum number of cycles N _cmax , the number of ants m′, the initialization time slice t=0, the cycle control variable N _c =0, the ant cycle variable k=0, the tabu table tabu _k (k=1, 2, ..., m'), node dynamic adjustment table tabv _k (k=1, 2,..., m'), path table path _k (k=1, 2,..., m'), place m' ants in m On a starting node, make the pheromone τ _ij (t)=const on the connected arcs of different nodes, where const is a constant; put the nodes in all destination node sets into the node dynamic adjustment table tabv _k (k=1, 2 ,..., m'); the ant is a variable used for optimization in the ant colony algorithm;

Step 2: The number of cycles N _c =N _c+1 , if Nc>N _cmax , exit the cycle and jump to step 11;

Step 3: The number of ants k=k+1, if k>m, exit the k loop and jump to step 8;

Step 4: Determine the node where the ant is currently located, and detect whether the node belongs to the destination node set, if so, delete the destination node corresponding to the node from the node dynamic adjustment table tabv _k , otherwise jump to the next step;

Step 5: Find the path connected to the node where the ant is located, calculate the transition probability according to the state transition probability formula, and select the next node starting from this node;

Step 6: Modify the tabu table _tabuk , put the newly selected node into the tabu table; and update the pheromone;

Step 7: If the tabu table tabu _k already contains all the nodes in the set of starting nodes and the set of destination nodes, then the ant's current route selection is over, and jump to step 9, otherwise, jump to step 5;

Step 8: Calculate and record the sum of new pheromones added by the ants in this round of node selection. If the route selection of all ants in this round is over, go to step 9, otherwise go to step 3;

Step 9: Compare the sum of the newly added pheromones of the combined computing nodes selected by all ants in this round, take the combined sequence of computing nodes with the largest sum of newly added pheromones, and perform pheromone update on the pheromones on the corresponding path according to the pheromone update rules. update;

Step 10: Clear tabu table tabu _k and path table path _k , and jump to step 2;

Step 11: Take the deployment route represented by the maximum pheromone path as the optimal solution of the deployment scheme.

It is also possible to apply real-time perception information to the best logistics deployment plan that has been generated to determine whether the plan needs to be adjusted dynamically. Delete the corresponding path in the optimal scheduling path; for the current scheduling task to be deleted, added, or modified, delete or add the corresponding node in the starting node set and destination node set; determine the current node position of the dispatching vehicle or the upcoming Arrived at the node position, after updating the latest state, generate the corresponding node state graph; place all k ants on the node at the current position of the dispatching vehicle, and re-optimize the solution.

The logistics allocation implementation device of the present invention may include a mathematical model generation module and a logistics allocation module; wherein,

The mathematical model generation module is used to generate a mathematical model for realizing logistics deployment through modeling of logistics deployment constraints, modeling of distances from deployment stations, and modeling of logistics deployment algorithms;

The logistics allocation module is used for applying the ant colony algorithm to allocate logistics on the basis of the mathematical model.

The mathematical model generation module may include a logistics allocation constraint modeling unit for:

Assume that the logistics allocation system has n logistics allocation stations in total, and each allocation station is denoted as a _i (i=1...n). At the future time T(t), the inventory of certain logistics goods in the allocation station a _i is S _i ( T(t)), and at this moment, the deployment station’s demand for the logistics goods is y _i (T(t)), which satisfies the state constraint condition S _i (T(t))≥y _i (T(t )); (i=1...n).

The logistics allocation constraint modeling unit can also be used for:

Logistics allocation management is carried out by establishing and maintaining a finite state machine for each logistics allocation station; the state machine function of logistics allocation station a _i is defined as: St _i (T(t)); where (i=1...n );

When the state machine cannot satisfy the state constraint condition at the corresponding time point in the future, the state transition of the state machine is performed.

The mathematical model generation module may include a distribution station distance modeling unit for absolute distance modeling and relative distance modeling; wherein,

In absolute distance modeling, set the distance between site a _i and site a _j as d _ij , and establish a two-dimensional matrix of distances between different sites:

When modeling relative distance, if there are no other stations on the path between station a _i and station a _j , then set the distance between station a _i and station a _j as d _ij ; if the distance between station a _i and station a _j There is a third station k in between, then d _ij =d _ik +d _kj ;

By analogy, the distance between all different sites is expressed as the sum of the indecomposable pairwise minimum distances between sites:

Among them, d′ _ij represents the relative distance between site a _i and site a _j , expressed as the relative value of the minimum distance between two sites.

The mathematical model generation module may include a logistics allocation algorithm modeling unit for predicting the future state, including the definition of a future state function and the definition of a logistics allocation matrix; wherein,

When defining the future state function, the future state prediction function is defined as: Num _i = fc(st _i (T(t)), y _i (T(t)), S _i (T(t)));

If Num _i > 0, it means that the allocation station a _i has remaining logistics goods, and the number is Num _i ; if Num _i = 0, it means that the allocation station a _i has no remaining logistics goods; when Num _i ≥ 0, it means that the allocation station a _i does not The logistics goods need to be supplemented through the logistics allocation program; if Num _i <0, it means that the allocation station a _i needs to start the logistics allocation program, and the number of logistics goods allocated to the allocation station a _i is the absolute value of Num _i ;

When defining the logistics allocation matrix, establish a two-dimensional logistics allocation matrix between the allocation stations. If the allocation station a _i does not need to allocate goods to the allocation station a _j , it is represented by 0; if it needs to be allocated, it is expressed as the value of the allocated logistics goods t _ij :

The logistics allocation algorithm modeling unit can also be used for:

use Indicates the sum of all logistics goods shipped to the deployment station a _j , which needs to meet the conditions:

That is, the logistics goods transported to the deployment station a _j must meet the requirements of the future state machine.

use Indicates the sum of the logistics goods transported from the deployment station a _i to all other deployment stations, and the conditions must be met: That is, the sum of the logistics goods transported by the deployment station a _i to all other deployment stations must meet the future state machine constraints of the deployment station a _i .

When the logistics allocation module applies the ant colony algorithm to allocate logistics on the basis of the mathematical model, the following steps can be performed:

Step 1: Set the maximum number of cycles N _cmax , the number of ants m′, the initialization time slice t=0, the cycle control variable N _c =0, the ant cycle variable k=0, the tabu table tabu _k (k=1, 2, ..., m'), node dynamic adjustment table tabv _k (k=1, 2,..., m'), path table path _k (k=1, 2,..., m'), place m' ants in m On a starting node, make the pheromone τ _ij (t)=const on the connected arcs of different nodes, where const is a constant; put the nodes in all destination node sets into the node dynamic adjustment table tabv _k (k=1, 2 ,..., m'); the ant is a variable used for optimization in the ant colony algorithm;

Step 2: Number of cycles N _c =N _c+1 , if N _c >N _cmax , exit the cycle and jump to step 11;

Step 3: The number of ants k=k+1, if k>m, exit the k loop and jump to step 8;

Step 4: Determine the node where the ant is currently located, and detect whether the node belongs to the destination node set, if so, delete the destination node corresponding to the node from the node dynamic adjustment table tabv _k , otherwise jump to the next step;

Step 5: Find the path connected to the node where the ant is located, calculate the transition probability according to the state transition probability formula, and select the next node starting from this node;

Step 6: Modify the tabu table _tabuk , put the newly selected node into the tabu table; and update the pheromone;

Step 7: If the tabu table tabu _k already contains all the nodes in the set of starting nodes and the set of destination nodes, then the ant's current route selection is over, and jump to step 9, otherwise, jump to step 5;

Step 8: Calculate and record the sum of new pheromones added by the ants in this round of node selection. If the route selection of all ants in this round is over, go to step 9, otherwise go to step 3;

Step 9: Compare the sum of the newly added pheromones of the combined computing nodes selected by all ants in this round, take the combined sequence of computing nodes with the largest sum of newly added pheromones, and perform pheromone update on the pheromones on the corresponding path according to the pheromone update rules. update;

Step 10: Clear tabu table tabu _k and path table path _k , and jump to step 2;

Step 11: Take the deployment route represented by the maximum pheromone path as the optimal solution of the deployment scheme.

The logistics allocation module can also correspond to the optimal logistics allocation plan that has been generated, apply real-time perception information, determine the execution degree of the plan, and delete corresponding tasks from the starting node set and destination node set for tasks that have been allocated. Node, delete the corresponding path from the optimal scheduling path; for the current scheduling task to be deleted, added, or modified, delete or add the corresponding node in the starting node set and destination node set; determine the node where the dispatching vehicle is currently located The position or the node position that is about to arrive, after the latest state update, generate the corresponding node state diagram; place all k ants on the current position node of the dispatching vehicle, and re-optimize the solution.

To sum up, it can be seen that, whether it is a method or a device for realizing the method, the technology for realizing logistics deployment in the present invention can generate logistics deployment constraints modeling, distribution station distance modeling, and logistics deployment algorithm modeling to realize logistics deployment. The mathematical model; and on the basis of the mathematical model, the ant colony algorithm is used to allocate logistics, so it can adapt to the dynamic changes in the logistics allocation process, and generate a nearly optimal logistics allocation plan in the least amount and within the smallest range, which can Improve the efficiency of logistics allocation to a large extent, achieve high-efficiency logistics allocation, and save logistics allocation costs.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention.

Claims (10)

1. A method for realizing logistics deployment is characterized by comprising the following steps:

generating a mathematical model for realizing logistics allocation through logistics allocation constraint condition modeling, allocation station distance modeling and logistics allocation algorithm modeling; applying an ant colony algorithm to allocate logistics on the basis of the mathematical model;

the step of applying the ant colony algorithm to allocate the logistics on the basis of the mathematical model comprises the following steps: on the basis of the mathematical model, an ant colony algorithm is applied to obtain the optimal scheduling scheme of the scheduling; the method further comprises the following steps: when the optimal scheduling scheme of the current scheduling is executed, the ant colony algorithm is applied to adjust the optimal scheduling scheme of the current scheduling;

the logistics allocation constraint condition modeling method comprises the following steps:

let the logistics dispatching system have n total logistics dispatching stations, each dispatching station is denoted as a_i(i-1 … n), at time t (t), the dispatching station a_iHas a certain logistics cargo inventory of S_i(T (t)), and at this time the demand of the distribution station for the physical distribution goods is y_i(T (t)) satisfying a state constraint condition S_i(T(t))≥y_i(T(t))；(i＝1…n)；

The allocation station distance modeling comprises absolute distance modeling and relative distance modeling; wherein, when modeling absolute distance, setting a site a_iAnd site a_jA distance d between_ijEstablishing a two-dimensional matrix of distances between different stations:

when modeling relative distance, if site a_iTo site a_jIf there is no other station on the path between the two, then set station a_iTo site a_jA distance d between_ij(ii) a If site a_iTo site a_jThere is a third site k in between, then d'_ij＝d_ik+d_kj(ii) a By analogy, the distance between all different sites is expressed as the sum of two non-resolvable minimum inter-site distances:

wherein, d'_ijIndicating site a_iTo site a_jThe relative distance between the two stations is expressed as the relative value of the distance between every two minimum stations;

the logistics allocation algorithm modeling is used for predicting a future state, and comprises a future state function definition and a logistics allocation matrix definition; wherein, go intoWhen defining the future state function, the future state prediction function is defined as: num_i＝fc(st_i(T(t)),y_i(T(t)),S_i(T (t)); wherein S is_i(T (t)) represents the inventory of goods in a certain stream at a certain dispatching station at a certain future time T (t), y_i(T (t)) represents the demand for said certain logistics goods at said T (t) said certain dispatching station, S_i(T), (t) and y_i(T (t)) satisfies the State constraint S_i(T(t))≥y_i(T(t))；St_i(t)) is a state machine function of said certain flow allocation station; fc () function is S_i(T(t))-y_i(T), (t)), and (st) is simultaneously performed_i(t)) satisfies the state constraint at the time t (t); wherein (i ═ 1 … n); if Num_i>0, denotes the deployment station a_iWith the remaining logistics goods in Num number_i(ii) a If Num_i0 denotes the order station a_iNo logistics goods remain; num_iWhen the value is more than or equal to 0, the value indicates a blending station a_iThe logistics goods do not need to be supplemented through a logistics allocation program; if Num_i<0, denotes the deployment station a_iStarting a logistics deployment program and deploying to a deployment station a_iThe number of the logistics goods is Num_iAbsolute value of (d); when defining the logistics allocation matrix, establishing a logistics allocation two-dimensional matrix between allocation stations, if the allocation station a_iWithout going to the allocating station a_jWhen the goods are prepared, 0 is used for representing; if the allocation is needed, the value is expressed as the allocated logistics cargo value t_ij：

2. The method of claim 1, further comprising:

the management of logistics deployment is carried out by establishing and maintaining a finite state machine for each logistics deployment station; logistics allocation station a_iThe state machine function of (a) is defined as: st_i(T (t)); wherein (i ═ 1 … n);

and when the state machine can not meet the state constraint condition of the corresponding future time point, the state machine is subjected to state conversion by starting a logistics allocation scheme so as to meet the state constraint condition.

3. The method of claim 1, further comprising:

by usingIndicating transport to dispatch station a_jThe sum of all the logistics cargo of (a) needs to satisfy the condition:

transported to the allocating station a_jThe logistics goods need to meet the requirements of a future state machine;

by usingIndicating a slave deployment station a_iThe total of logistics goods transported to all other allocation stations needs to meet the following conditions:namely the dispatching station a_iThe total sum of the logistics goods transported to all other allocation stations needs to meet the requirement of the allocation station a_iFuture state machine constraints.

4. The method according to any one of claims 1 to 3, wherein the process of formulating the logistics using the ant colony algorithm based on the mathematical model comprises the steps of:

step 1: setting a maximum number of cycles N_cmaxNumber of ants m', initialization time slice t is 0, and loop control variable N_c0, 0 for ant circulation variable k, tabu_k(k ═ 1,2, …, m'), node dynamic adjustment table tabv_k(k ═ 1,2, …, m'), pathsTable path_k(k ═ 1,2, …, m '), placing m ' ants on m ' starting nodes, making different nodes connected to pheromone tau on arc_ij(t) const, wherein const is a constant; putting nodes in all destination node sets into a node dynamic adjustment table tabv_k(k ═ 1,2, …, m'); the ants are variables used for optimizing in the ant colony algorithm;

step 2: number of cycles N_c＝N_c+1If N is present_c＞N_cmaxExiting the loop and jumping to step 11;

and step 3: if k is larger than m, quitting k circulation and jumping to step 8;

and 4, step 4: determining the node where the ant is located at present, detecting whether the node belongs to a destination node set, and if so, dynamically adjusting a destination node slave node table tabv corresponding to the node_kDeleting, otherwise, skipping to the next step;

and 5: searching a path connected with a node where the ant is located, calculating a transition probability according to a state transition probability formula, and selecting a next node from the node;

step 6: tabu modification_kPutting the newly selected node into a tabu table; and updating pheromones;

and 7: if tabu is contraindicated_kIf all the nodes in all the starting node set and all the nodes in the target node set are included, the ant finishes the route selection, and the step 9 is skipped, otherwise, the step 5 is skipped;

and 8: calculating and recording the sum of newly added pheromones of the ants in the round of selection nodes, if all ants in the round are finished selecting the route, skipping to the step 9, otherwise skipping to the step 3;

and step 9: comparing the newly increased pheromone sum of the combined calculation nodes selected by all ants in the round, taking the calculation node combination sequence with the maximum newly increased pheromone sum value, and updating pheromones on corresponding paths according to pheromone updating rules;

step 10: tabu for emptying tabu table_kPath table path_kSkipping to the step 2;

step 11: and taking the deployment route represented by the maximum pheromone path as the optimal solution of the deployment scheme.

5. The method of claim 4, further comprising:

corresponding to the generated optimal logistics allocation scheme, determining whether the scheme needs to be dynamically adjusted or not by applying real-time perception information, deleting corresponding nodes from the starting node set and the target node set and deleting corresponding paths from the optimal scheduling paths for the allocated tasks; deleting and adding corresponding nodes in a starting node set and a target node set for the scheduling tasks to be deleted, added and modified currently; determining the current node position of the dispatching vehicle or the node position to be reached, and generating a corresponding node state diagram after updating the latest state; and (5) placing all k ants at the current position node of the dispatching car, and carrying out optimization solution again.

6. A realization device for logistics allocation is characterized by comprising a mathematical model generation module and a logistics allocation module; wherein,

the mathematical model generation module is used for generating a mathematical model for realizing logistics distribution through logistics distribution constraint condition modeling, distribution station distance modeling and logistics distribution algorithm modeling;

the logistics allocation module is used for allocating logistics by applying an ant colony algorithm on the basis of the mathematical model; the step of applying the ant colony algorithm to allocate the logistics on the basis of the mathematical model comprises the following steps: on the basis of the mathematical model, an ant colony algorithm is applied to obtain the optimal scheduling scheme of the scheduling; when the optimal scheduling scheme of the current scheduling is executed, the ant colony algorithm is applied to adjust the optimal scheduling scheme of the current scheduling;

the mathematical model generation module comprises a logistics allocation constraint condition modeling unit and is used for:

let the logistics dispatching system have n total logistics dispatching stations, each dispatching station is denoted as a_i(i-1 … n) and at a future time t (t), the dispatching station a_iHas a certain logistics cargo inventory of S_i(T (t)), and at this time the demand of the distribution station for the physical distribution goods is y_i(T (t)) satisfying a state constraint condition S_i(T(t))≥y_i(T(t))；(i＝1…n)；

The mathematical model generation module comprises a distribution station distance modeling unit and is used for carrying out absolute distance modeling and relative distance modeling; wherein, when modeling absolute distance, setting a site a_iAnd site a_jA distance d between_ijEstablishing a two-dimensional matrix of distances between different stations:

when modeling relative distance, if site a_iTo site a_jIf there is no other station on the path between the two, then set station a_iTo site a_jA distance d between_ij(ii) a If site a_iTo site a_jThere is a third site k in between, then d'_ij＝d_ik+d_kj(ii) a By analogy, the distance between all different sites is expressed as the sum of two non-resolvable minimum inter-site distances:

wherein, d'_ijIndicating site a_iTo site a_jThe relative distance between the two stations is expressed as the relative value of the distance between every two minimum stations;

the mathematical model generation module comprises a logistics allocation algorithm modeling unit which is used for predicting future states, including future state function definition and logistics allocation matrix definition; when defining the future state function, the future state prediction function is defined as: num_i＝fc(st_i(T(t)),y_i(T(t)),S_i(T (t)); wherein S is_i(T (t)) represents the inventory of goods in a certain logistics at a certain deployment site at a certain future time T (t),y_i(T (t)) represents the demand for said certain logistics goods at said T (t) said certain dispatching station, S_i(T), (t) and y_i(T (t)) satisfies the State constraint S_i(T(t))≥y_i(T(t))；St_i(t)) is a state machine function of said certain flow allocation station; fc () function is S_i(T(t))-y_i(T), (t)), and (st) is simultaneously performed_i(t)) satisfies the state constraint at the time t (t); wherein (i ═ 1 … n); if Num_i>0, denotes the deployment station a_iWith the remaining logistics goods in Num number_i(ii) a If Num_i0 denotes the order station a_iNo logistics goods remain; num_iWhen the value is more than or equal to 0, the value indicates a blending station a_iThe logistics goods do not need to be supplemented through a logistics allocation program; if Num_i<0, denotes the deployment station a_iStarting a logistics deployment program and deploying to a deployment station a_iThe number of the logistics goods is Num_iAbsolute value of (d); when defining the logistics allocation matrix, establishing a logistics allocation two-dimensional matrix between allocation stations, if the allocation station a_iWithout going to the allocating station a_jWhen the goods are prepared, 0 is used for representing; if the allocation is needed, the value is expressed as the allocated logistics cargo value t_ij：

7. The apparatus of claim 6, wherein the logistics deployment constraint modeling unit is further configured to:

the management of logistics deployment is carried out by establishing and maintaining a finite state machine for each logistics deployment station; logistics allocation station a_iThe state machine function of (a) is defined as: st_i(T (t)); wherein (i ═ 1 … n);

and when the state machine can not meet the state constraint condition of the corresponding future time point, carrying out state conversion on the state machine.

8. The apparatus of claim 6, wherein the logistics deployment algorithm modeling unit is further configured to:

by usingIndicating transport to dispatch station a_jThe sum of all the logistics cargo of (a) needs to satisfy the condition:

transported to the allocating station a_jThe logistics goods need to meet the requirements of a future state machine;

by usingIndicating a slave deployment station a_iThe total of logistics goods transported to all other allocation stations needs to meet the following conditions:namely the dispatching station a_iThe total sum of the logistics goods transported to all other allocation stations needs to meet the requirement of the allocation station a_iFuture state machine constraints.

9. The apparatus according to any one of claims 6 to 8, wherein the logistics allocation module is configured to perform the following steps when applying the ant colony algorithm to allocate logistics based on the mathematical model:

step 1: setting a maximum number of cycles N_cmaxNumber of ants m', initialization time slice t is 0, and loop control variable N_c0, 0 for ant circulation variable k, tabu_k(k ═ 1,2, …, m'), node dynamic adjustment table tabv_k(k ═ 1,2, …, m'), path table path_k(k ═ 1,2, …, m '), placing m ' ants on m ' starting nodes, making different nodes connected to pheromone tau on arc_ij(t) const, wherein const is a constant; putting nodes in all destination node sets into node dynamicsTabv of regulation table_k(k ═ 1,2, …, m'); the ants are variables used for optimizing in the ant colony algorithm;

step 2: number of cycles N_c＝N_c+1If N is present_c＞N_cmaxExiting the loop and jumping to step 11;

and step 3: if k is larger than m, quitting k circulation and jumping to step 8;

and 4, step 4: determining the node where the ant is located at present, detecting whether the node belongs to a destination node set, and if so, dynamically adjusting a destination node slave node table tabv corresponding to the node_kDeleting, otherwise, skipping to the next step;

and 5: searching a path connected with a node where the ant is located, calculating a transition probability according to a state transition probability formula, and selecting a next node from the node;

step 6: tabu modification_kPutting the newly selected node into a tabu table; and updating pheromones;

and 7: if tabu is contraindicated_kIf all the nodes in all the starting node set and all the nodes in the target node set are included, the ant finishes the route selection, and the step 9 is skipped, otherwise, the step 5 is skipped;

and 8: calculating and recording the sum of newly added pheromones of the ants in the round of selection nodes, if all ants in the round are finished selecting the route, skipping to the step 9, otherwise skipping to the step 3;

and step 9: comparing the newly increased pheromone sum of the combined calculation nodes selected by all ants in the round, taking the calculation node combination sequence with the maximum newly increased pheromone sum value, and updating pheromones on corresponding paths according to pheromone updating rules;

step 10: tabu for emptying tabu table_kPath table path_kSkipping to the step 2;

step 11: and taking the deployment route represented by the maximum pheromone path as the optimal solution of the deployment scheme.

10. The apparatus of claim 9, wherein the logistics deployment module is further configured to:

corresponding to the generated optimal logistics allocation scheme, determining the execution degree of the scheme by using real-time perception information, deleting corresponding nodes from the starting node set and the destination node set and deleting corresponding paths from the optimal scheduling paths for the allocated tasks; deleting and adding corresponding nodes in a starting node set and a target node set for the scheduling tasks to be deleted, added and modified currently; determining the current node position of the dispatching vehicle or the node position to be reached, and generating a corresponding node state diagram after updating the latest state; and (5) placing all k ants at the current position node of the dispatching car, and carrying out optimization solution again.