CN115081674B

CN115081674B - Local container transportation typesetting optimization method under novel truck queuing driving mode

Info

Publication number: CN115081674B
Application number: CN202210498853.7A
Authority: CN
Inventors: 许敏
Original assignee: Shenzhen Research Institute HKPU
Current assignee: Shenzhen Research Institute HKPU
Priority date: 2022-05-09
Filing date: 2022-05-09
Publication date: 2023-04-18
Anticipated expiration: 2042-05-09
Also published as: CN115081674A

Abstract

The invention discloses a novel local container transportation typesetting optimization method under a truck queuing running mode, which comprises the following steps: acquiring a network map; determining an objective function of the total operation cost according to the network diagram; wherein the total operating cost comprises: driver hiring costs, truck deployment costs, total fuel consumption costs for trucks to allow for queuing fuel savings, and travel costs resulting from alternative modes of transportation; determining an initial solution corresponding to the network diagram; and under the constraint condition, updating the initial solution to minimize the objective function to obtain an updated solution corresponding to the network diagram. The invention can make the optimal hiring quantity, service route and dispatching plan of drivers and trucks for large-scale container hauling tasks in a short time, thereby improving the utilization rate of the trucks, reducing the quantity of the trucks and reducing the total operation cost.

Description

Local container transportation typesetting optimization method under novel truck queuing driving mode

Technical Field

The invention relates to the technical field of operation and transportation of land vehicles running on a non-rail, in particular to a method for optimizing transportation and layout of local containers in a novel truck queuing running mode.

Background

Global trade relies heavily on container transport where intermodal containers are moved by ship, rail and truck between seaports, docks and local customers (DHL, 2021). Throughout the transportation chain, the carrier provides delivery and pickup services to the recipient and the shipper, respectively, through short-haul container transportation by truck between the terminal and the local customer, which is known as the local container transport problem (LCDP). Although the distance of transportation is very short compared to ocean and land-based long haul, the operating cost of LCDP is approximately as high as 80% of the total cost of the entire container chain (Macharis and bonteking, 2004). Thus, the cost effectiveness of local container transport services is particularly important to the profitability of multimodal intermodal container transport. In the traditional container hauling mode, each consignee/shipper needs a driver to leave the dock with a full/empty truck and wait for the unloading/loading work at the customer site; after the unloading/loading work is completed, the driver should return the empty/loaded truck to the dock. In recent decades, in order to reduce the overall operating costs, the pallet mode has been improved in many respects, such as sharing of empty trucks, where empty trucks off the consignee can be reused by the shipper without returning to the dock (Imai et al, 2007). With the development of networking and autonomous vehicle (CAV) technology, emerging truck-in-line driving models help reduce labor in LCDP. In this mode, a group of autonomous cars are driven together at small intervals, and only the leading truck requires the driver to drive. In addition, many field trials found that trucks traveling in line could also reduce fuel consumption by reducing aerodynamic drag, especially for following trucks, this model provides a significant fuel savings of about 10-20% (Davila et al, 2013; lammert et al, 2014; lu and Shladover, 2014). How to take advantage of emerging technologies to gain the greatest cost effectiveness is one of the major issues facing container transport service providers.

In the past decades, many scholars have conducted relevant research on LCDP, taking into account various practical constraints and characteristics, such as resource constraints, flexible orders, and uncertainty in travel speed. For example, zhang considers the limited number of trucks and empty containers available at the dock. They built a Mixed Integer Programming (MIP) model on the directed graph and proposed an algorithm based on reactive tabu search to solve this problem. Escudero studies the daily journey problem of taking into account the uncertainty in journey time due to accidents. They developed a real-time optimization model and proposed a genetic algorithm for fleet dynamic planning and management. Zhan establishes a mixed integer nonlinear programming model for dynamic container transport services with flexible container orders and develops four different solution methods to solve the problem. Benantar solves the real-world problem of container transport in view of the availability of containers. There have also been many studies on LCDP in a new, resource-efficient, cost-effective mode of transportation. For example, imai proposes an LCDP that considers empty car sharing and develops a heuristic algorithm based on lagrangian relaxation to obtain a near-optimal solution by solving several sub-problems. Thereafter, to further improve shipping efficiency, xue proposes a lead truck-following truck separation mode, in which the lead truck can be separated from its mating following truck, enabling the driver to continue servicing the next customer without waiting for unloading/loading work at the current customer site. They describe this problem as a vehicle path problem with time constraints and propose a tabu search algorithm. Song considers the necessary container maintenance of the warehouse on the basis of Xue. They treat the problem as an asymmetric vehicle path problem and formulate it as an MILP model. Aiming at the problem, a branch pricing and cutting algorithm is provided. Both Song and Xue assume that a tractor can only carry one trailer at a time. Later, some studies have relaxed this assumption, allowing tractors to tow two or more trailers (containers) at a time. In particular, zhang et al explored a collapsible container transport service in which trucks could transport one loaded container or multiple collapsible containers at a time. To address this problem, they propose a mileage-based truck state transition method and an improved reactive tabu search algorithm. Wang and Zhang propose a dual trailer split mode of LCDP in which one tractor can carry two trailers. Zhang et al further extends it to the multi-trailer pull-down mode, where a mixed integer nonlinear programming model and a backtracking adaptive threshold acceptance algorithm were developed to solve the problem.

In the prior art, vehicles/trucks are queued for great potential to further reduce costs and increase efficiency during transportation via CAV technology, and there have been some studies to incorporate the emerging truck-queuing technology into LCDP. In these studies, you et al pioneered a generic LCDP that allows for sharing empty trucks in a truck-in-line mode of operation where only the lead truck requires the driver's drive and is automatically tracked by a group of unmanned trucks. They also assume that all drivers and trucks must return to the dock after all delivery/pickup requests from the customer have been completed within the specified driver's maximum daily hours of operation, and propose a heuristic solution based on the ant colony algorithm. Xue et al later incorporated the fuel savings benefits of truck rows into a special truck-in-line driving based LCDP model where different rows of trucks could not be queued together throughout the service and sharing of empty trucks was not allowed. Recently, you et al considered the load-related fuel costs in LCDP and the multi-trip of the driver and the in-line driving of the truck, where the driver could enter the dock multiple times during a work day, but did not allow sharing of empty trucks among customers. They developed a MIP model and proposed a custom branch and price cut algorithm to solve this problem. However, the studies by Xue et al, you et al, and You et al assume that an unmanned truck left empty at the customer site will not be available for other service tasks before regaining driver or lead truck instructions, which may result in low truck utilization and higher overall operating costs.

Accordingly, the prior art is yet to be improved and developed.

Disclosure of Invention

The technical problem to be solved by the invention is to provide a novel optimization method for local container transportation typesetting in a truck queuing running mode aiming at the defects in the prior art, and the method aims at solving the problem that the utilization rate of trucks is low in the prior art, so that the total operation cost is high.

The technical scheme adopted by the invention for solving the technical problem is as follows:

a method for optimizing local container transportation typesetting in a novel truck queuing driving mode comprises the following steps:

acquiring a network map; wherein the network graph comprises a set of nodes and a set of arcs, the nodes in the set of nodes comprise a delivery client node, a pickup client node and a dock, and the arcs in the set of arcs are connecting arcs between two nodes; the delivery client node comprises: the task nodes before unloading and the task nodes after unloading, wherein the goods-taking client node comprises: the task nodes before loading and the task nodes after loading;

determining an objective function of the total operation cost according to the network graph; wherein the total operating cost comprises: driver hiring costs, truck deployment costs, total truck fuel consumption costs to allow for queuing fuel savings, and travel costs resulting from alternative modes of transportation;

determining an initial solution corresponding to the network diagram;

and under the constraint condition, updating the initial solution to minimize the objective function to obtain an updated solution corresponding to the network diagram.

The method for optimizing the local container transportation typesetting in the novel truck queuing driving mode comprises the following steps:

where min represents a minimization function, x represents a binary route decision variable for the driver, y represents a binary route decision variable for the truck, α represents a binary decision variable for the truck to traverse the arc, β represents a binary decision variable for the truck to traverse over the arc, s represents a time variable, λ ₁ Represents the driver's daily fixed cost, V represents the driver, V represents the set of drivers, j represents a task node, C represents the set of task nodes other than the dock,

binary route decision variable representing driver v on arc (0, j), 0 representing dock, λ ₂ Represents the daily fixed cost score of the truck, K represents the collection of trucks, and->

Binary route decision variable, c, representing truck k on arc (0, j) _ij Representing the fuel consumption cost of a truck on arc (i, j), i representing the mission node, N representing the set of nodes comprising the quay, β _ij Binary decision variables indicating that there is a truck crossing on arc (i, j), η indicates the fuel savings rate, based on the trucks in the truck fleet column>

Indicating truck k is in arc (i)Binary route decision variable on j), c' _ij Represents the travel cost on arc (i, j) instead of the mode of transportation, based on the number of hours on the trip>

Represents the binary route decision variable for driver v on arc (i, j), Σ represents the sum, and e represents the membership.

The novel local container transportation typesetting optimization method under the truck queuing driving mode is characterized in that the constraint conditions are as follows:

/>

wherein, the first and the second end of the pipe are connected with each other,

a binary decision variable indicating whether or not an arc (i, D) is crossed in a trip connecting OD pairs (0, D), D indicating a task node of a delivery task, D ₁ Indicating ren before unloadingService node collection>

A binary decision variable indicating whether or not an arc (d, j) is crossed during a trip connecting the OD pair (d, p), p indicating the task node of the pick task, and/or->

A binary decision variable representing whether or not to cross an arc (D, j) in a run connecting an OD pair (D, 0), N \ D represents a set of N minus D, D ₂ Means for indicating a set of task nodes on dump>

A binary decision variable representing whether or not an arc (i, p) is crossed during a trip connecting the OD pair (d, p), and->

Binary decision variables representing whether or not to cross the arc (i, P) in the run connecting the OD pair (0, P), N \ P represents the set of N minus P, P ₁ Represents a collection of pre-loaded task nodes, and>

a binary decision variable, P, representing whether or not to cross the arc (P, j) in the run connecting the OD pair (P, 0) ₂ Represents a set of loaded task nodes, based on the status of the task node, and/or the status of the task node>

Binary decision variable indicating whether or not an arc (0, i) is crossed in a trip connecting OD pairs (o, d), o indicating the node at which the task starts, R indicating the set of OD pairs, </R >>

A binary decision variable representing whether or not an arc (i, 0) is crossed during a trip connecting the OD pair (o, d), and->

Binary decision representing whether or not to cross arc (i, o) in the course of connecting OD pair (o, d)Variable N \ o represents the set of N minus o>

A binary decision variable indicating whether or not an arc (d, i) is crossed during a trip connecting the OD pair (0, d), N \ d indicating the set of N minus d, and/or>

Indicates presence, or is present>

A binary route decision variable representing driver v on arc (0, i), N \ j representing the set of N minus j, and/or>

A binary route decision variable representing driver v on arc (j, i), based on the decision value of the driver v>

Represents a binary route decision variable, C, for truck k over arc (0, i) ₁ Represents D ₁ And P ₁ In the same client node, i' represents a task node in the same client node that corresponds to task node i, and +>

Represents a binary route decision variable for truck k on arc (j, i), N \ j, j ' } represents a set of N minus j and j ', j ' represents a task node corresponding to task node j in the same client node, L represents the number of largest trucks in the truck queue, and/or>

A binary decision variable representing whether or not an arc (i, g) is crossed during a trip connecting the OD pair (o, d), and->

A binary decision variable indicating whether or not an arc (g, j) is crossed in a trip connecting an OD pair (o, d), g indicating a task node other than o and d in C, g' indicating a task node in the same client node as the task nodeg corresponding task node, C \ g, g '} denotes C removes the set of g and g'; and holds>

A binary decision variable representing whether or not an arc (o, i) is crossed during a trip connecting the OD pair (o, d), and->

Binary decision variable representing whether or not an arc (j, d) is crossed in a trip connecting an OD pair (o, d), based on a decision value>

Binary decision variable, s, representing whether or not to cross arc (i, j) in the run connecting OD pairs (o, d) _i Representing the service start time, p, of the task node i _i Indicating the time of discharge/loading, s, of the task node i _i′ Representing the service start time, t, of the task node i _0i Representing the travel time, β, of the driver by truck transfer from dock 0 to the service node i _0i Binary decision variable, t ', representing the passing of a truck over arc (0,i)' _0i Representing the time, t, of the driver's journey through the alternate mode of transportation on arc (0, i) _i0 Represents the time the driver travels by truck over arc (i, 0), T represents the maximum work time per day, C ₂ Is shown by D ₂ And P ₂ Union of (1), s _j Represents the service start time, t, of task node j _ij Representing the time, t ', of the driver's trip through the truck on arc (i, j) ' _ij Represents the travel time of the driver through the alternate mode of transportation on arc (i, j), and M represents a constant.

The method for optimizing the local container transportation typesetting in the truck queuing mode comprises the following steps of:

constructing a character string for a driver, and sequentially inserting task nodes which are not served from a wharf; when a task node which is not yet served is inserted, selecting a task node which is closest to the last inserted task node to insert;

ending the driver's string when the task node inserted in the driver's string reaches the maximum truck bank size constraint;

and constructing a character string for the next driver, and sequentially inserting the task nodes which are not served from the wharf until all the task nodes are inserted into the character string to obtain an initial solution corresponding to the network diagram.

The method for optimizing the local container transportation typesetting in the truck queuing mode comprises the following steps of updating the initial solution to minimize the objective function under the constraint condition to obtain an updated solution corresponding to the network diagram, wherein the updated solution comprises:

when the initial solution does not meet the constraint condition of task node time arrangement, re-determining the initial solution corresponding to the network graph;

when the initial solution meets the constraint condition of task node time arrangement, evaluating the initial solution based on the objective function to obtain an evaluation result;

and when the evaluation result meets the preset requirement, taking the initial solution as an updated solution corresponding to the network diagram.

The method for optimizing the local container transportation typesetting in the truck queuing mode comprises the following steps of updating the initial solution to minimize the objective function under the constraint condition to obtain an updated solution corresponding to the network diagram, and further comprises the following steps:

when the evaluation result does not meet the preset requirement, searching the initial solution by adopting a neighborhood search operator to obtain a neighborhood solution; wherein the neighborhood search operator comprises: an exchange operator, an inversion operator and an insertion operator;

when the objective function value corresponding to the domain solution is smaller than the objective function value corresponding to the initial solution, taking the neighborhood solution as the initial solution corresponding to the network graph, and reducing the temperature corresponding to the initial solution;

and when the initial solution meets the termination condition, taking the initial solution as an updated solution corresponding to the network graph.

when the objective function value corresponding to the domain solution is larger than or equal to the objective function value corresponding to the initial solution, determining the initial solution corresponding to the network graph according to a Metropolis criterion, and reducing the temperature corresponding to the initial solution;

and when the initial solution does not accord with the termination condition, continuously judging whether the initial solution meets the constraint condition of task node time arrangement or not, and when the initial solution meets the evaluation result and meets the preset requirement or the initial solution accords with the termination condition, taking the initial solution as an updating solution corresponding to the network graph.

The method for optimizing the local container transportation typesetting in the novel truck queuing mode comprises the following steps: the temperature corresponding to the initial solution is smaller than a preset temperature threshold value or the iteration number of the initial solution reaches the maximum continuous non-improvement iteration number;

and when the temperature corresponding to the initial solution is reduced, reducing the temperature according to the annealing rate.

A computer device comprising a memory storing a computer program and a processor, wherein the processor implements the steps of the method as claimed in any one of the above when executing the computer program.

A computer-readable storage medium, on which a computer program is stored, wherein the computer program, when being executed by a processor, carries out the steps of the method as set forth in any one of the preceding claims.

Has the advantages that: the invention can make the optimal hiring quantity, service route and dispatching plan of drivers and trucks for large-scale container hauling tasks in a short time, thereby improving the utilization rate of the trucks, reducing the quantity of the trucks and reducing the total operation cost.

Drawings

Fig. 1 is a schematic diagram of an example of application of the truck queue driving mode in the LCDP according to the present invention.

FIG. 2 is a heuristic construction algorithm HCSA framework of the present invention.

Fig. 3 is the change in CPU run time with increasing customer size.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention clearer and clearer, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

Referring to fig. 1-3, the present invention provides some embodiments of a method for optimizing local container transportation layout in a new truck queue driving mode.

As shown in fig. 1 to fig. 3, the method for optimizing local container transportation layout in the novel truck queue driving mode of the present invention includes the following steps:

s100, acquiring a network map; wherein the network graph comprises a set of nodes and a set of arcs, the nodes in the set of nodes comprise a delivery client node, a pickup client node and a dock, and the arcs in the set of arcs are connecting arcs between two nodes; the delivery client node comprises: the pre-unloading task node and the post-unloading task node, wherein the pick-up client node comprises: a task node before shipment and a task node after shipment.

Step S200, determining a target function of the total operation cost according to the network graph; wherein the total operating cost comprises: driver hiring costs, truck deployment costs, total fuel consumption costs for trucks to allow for queuing fuel savings, and travel costs resulting from alternative modes of transportation.

And step S300, determining an initial solution corresponding to the network diagram.

And S400, under the constraint condition, updating the initial solution to minimize the objective function to obtain an updated solution corresponding to the network diagram.

In particular, the trucks of the present application employ a train of trucks with at least one truck in the train, only the lead truck requiring the driver's driving, and a set of unmanned trucks that automatically track the lead truck. The number of trucks in the train of trucks is less than or equal to the maximum number of trucks. The trucks can be shared, that is to say, the leading truck of the unmanned truck can be adjusted as required, so that the utilization rate of the trucks is improved, the number of the trucks is reduced, and the total operation cost is reduced.

The existing research mainly aims at the optimization method of the local container transportation typesetting under the traditional truck queuing mode. The optimization method provided by the invention aims at a novel local container transportation mode, is more flexible and efficient compared with the existing mode, not only saves more manpower, material resources and financial resources, improves the service efficiency and reduces the total operation cost, but also can reduce the air resistance suffered by the truck when the truck lines up to run in a short distance, and effectively reduces the oil consumption and the pollutant emission of the truck. In addition, the patent provides a method for optimizing the local container problem in a novel transportation mode, so that the optimal employment number, service routes and scheduling plans of drivers and trucks can be set for large-scale container hauling tasks in a short time, and powerful technical support and guarantee are provided for managers or planners.

In order to reduce the total operation cost as much as possible, mathematical modeling is performed to form a set of points represented by N in a network graph G = (N, a), where a = { (i, j) | i, j ∈ N, i ≠ j } is a set of arcs. Note that N = douc P {0}, where D and P are the delivery and pickup customer node sets, and 0 represents a dock, with trucks being a homogeneous fleet of autonomous trucks of single-package capacity. In the new truck in-line mode of operation (IPOM), a driver may transfer from one customer node to another customer node via a truck or other mode of transportation, and may drive an unmanned truck at another customer node while transferring using other modes of transportation (alternate modes of transportation may be other than a truck, such as a motorcycle, a taxi, a shared vehicle, a taxi service, and a bus), to effect a rearrangement of unmanned trucks in a fleet of trucks. There will therefore be two travel times t associated with arc (i, j) ∈ A _ij And t' _ij Representing the time required to traverse arc (i, j) with trucks and other alternative vehicles, respectively. The fuel consumption cost of the truck and the travel cost of the alternative mode of transportation are respectively c _ij And c' _ij And (4) showing. During transport, trucks are driven in a queue to save fuel. And the number of trucks traveling together on each arc cannot exceed the maximum number of trucks L for safety reasons. Regarding the fuel saving effect, the fuel saving rate of trucks in the queue is assumed to be eta, and leading trucks cannot save fuel. For example, the total fuel cost produced by a queue of m trucks crossing arc (i, j) ∈ a would be calculated by the following constraint: c. C _ij [1+(1-η)(m-1)]. In addition, the daily fixed costs for the driver and truck are each lambda ₁ And λ ₂ And (4) showing.

In the IPOM mode, the driver can use other transportation means available nearby, go to other customer sites to perform service tasks, and drive the truck previously empty at the customer site back to the corresponding site, in addition to the truck. The operational characteristics of the IPOM are explained using the example shown in fig. 1. In this example, four delivery customers (including D1, D2, D3, and D4) and two pickup customers (i.e., P1 and P2) will be serviced. In fig. 1, the driver 1 is shown leaving the dock with a queue of three full trucks (driver 1 in the lead truck followed by 2 unmanned trucks). The driver 1 stops the train at D1, D2 and D3, respectively, separates 1 unmanned truck for unloading at D1 and D2, respectively, unloads the lead truck at D3, and the driver 1 waits at D3 until the unloading work is completed. Then, driver 1 again goes to D2, the lead truck and the unloaded 1 drone truck (i.e., empty truck) form a queue of two empty trucks, and the unloaded 1 drone truck (i.e., empty truck) at D1 is returned to the truck fleet of another driver (driver 2 in this example) and driver 2 is brought back to the dock. Since empty trucks may be shared among customers, driver 1 will give two empty trucks to pick-up customers P1 and P2, released by delivery customers D2 and D3, respectively, for reuse. After the truck queue of the driver 1 reaches the position P1, 1 unmanned empty truck is released, the leading truck is continuously driven to reach the position P2, after the loading work at the position P2 is finished, the driver 1 returns to the wharf with the full-load leading truck, and after the loading work of the unmanned empty truck at the position P1 is finished, the truck is conveyed back to the wharf by the driver 2, and at the moment, the truck is used as the leading truck. When the driver 2 leaves the dock, he takes a full truck and stops at the customer D4 for unloading. Since the IPOM allows the driver to move to P1 using other vehicles (such as taxis) after stopping all the lead trucks and drive back the full truck without waiting at D4 for the completion of the dump mission and then revisiting D4 to form a queue of two trucks. Since the driver 2 may have additional time available in the longest working hours after arriving again at D4, the driver 2 may continue to travel to D1, forming a queue of three trucks in total, and finally returning to the dock. However, if there is no IPOM, driver 2 must wait at D4 until the unloading work is completed, so driver 2 may not have enough time to complete the pickup at P1 without violating the maximum work time, and thus may be less likely to service D1. In this case, another driver or two are required to drive the trucks to bring the trucks at P1 and D1 back to the dock. As shown by the above examples, IPOM has great potential in improving transportation efficiency by saving labor and trucks.

The LCDP problem studied for the present invention, where empty trucks may come from a terminal or delivery customer who completed the unloading job. Furthermore, a queue mode where a plurality of trucks travel together in a queue will force the trucks to pass through a number of temporary customer points on the way from the respective origin to the respective target customer. Sharing empty trucks and truck queue travel patterns among customers presents challenges in modeling the specific travel path of the truck and calculating the number of trucks needed to complete all customer service. In fact, each customer is associated with two different tasks in both phases before and after the loading/unloading work. For modeling purposes, each customer node may be split into two task nodes to represent two phases of the customer task. In particular, with C ₁ The set of first stage task nodes is represented, including a delivery task node set (i.e., a set D of task nodes before unloading in a delivery task) requiring a fully loaded truck ₁ ) And a set of pick-up task nodes requiring empty trucks (i.e., pick-up tasks)Set of pre-shipment task nodes P ₁ ). Similarly, with C ₂ The set of second stage task nodes is represented, including a set of delivery task nodes for emptying trucks (i.e., a set D of unloaded task nodes in a delivery task) ₂ ) And a loaded pick-up task node set (i.e., a set P of loaded task nodes in a pick-up task) ₂ ). With C = C ₁ ∪C ₂ Representing the set of all task nodes except wharf, with C ₁ ＝D ₁ ∪P ₁ And C ₂ ＝D ₂ ∪P ₂ . The set of customer nodes N may then be converted to N = D ₁ ∪D ₂ ∪P ₁ ∪P ₂ U {0} = C ═ u {0}. Note that the first stage task nodes of the client i ∈ C ₁ The corresponding second stage task node is expressed as i' epsilon C ₂ And vice versa. According to the definition of a client node two-stage task node, the OD pairs can be defined as follows: the OD pair consisting of node o where a task starts and node d where a task ends is used to indicate that the truck from the starting point is dedicated to serving the end task node. For a particular OD pair, a set of task nodes that the truck sequentially traverses on a route from an origin to a destination may be considered a transit node. The wharf cannot act as a transport node in the path connecting the OD pair. Note that each task node can only act once as a start or end point, so the number of OD pairs is limited in the problem under investigation. The following constraints (1) - (8) are employed to track the truck movement trajectory in conjunction with the truck movement characteristics and the client node splitting strategy set forth above.

/>

a binary decision variable indicating whether or not an arc (i, D) is crossed in a trip connecting OD pairs (0, D), D indicating a task node of a delivery task, D ₁ Means for indicating a set of task nodes prior to offloading>

Binary decision variable which indicates whether or not an arc (d, j) is crossed in the course of a link OD pair (d, p), p indicating the task node of the pick task, and/or the device>

A binary decision variable representing whether or not to cross an arc (D, j) in a run connecting an OD pair (D, 0), N \ D represents a set of N minus D, D ₂ Represents a set of task nodes after a drop, and->

Binary decision variable representing whether or not to cross arc (i, p) in the course of connecting OD pair (d, p)Volume, or>

a binary decision variable, P, representing whether or not to cross the arc (P, j) in the course of connecting the OD pair (P, 0) ₂ Represents a set of loaded task nodes, based on the status of the task node, and/or the status of the task node>

A binary decision variable representing whether or not an arc (0, i) is crossed in a trip connecting an OD pair (o, d), o representing the node at which the task starts, C representing the set of nodes of the task other than the quay, R representing the set of OD pairs, and/or>

Binary decision variable indicating whether or not an arc (i, o) is crossed in a trip connecting OD pair (o, d), N \ o indicating the set of N minus o, based on>

Indicating the presence.

Constraints (1) - (4) represent truck movement requirements. More specifically, constraint (1) states that for the first phase delivery task node, a fully loaded truck must be transported from the dock to the node. Constraint (2) indicates that for the second stage delivery task node, the empty truck it released can be transported to the dock or the first stage pick task node for reuse; similarly, constraint (3) indicates that the demand for empty trucks for the first stage pick task node can be met by either the dock or the second stage delivery task node. Constraint (4) ensures that for the second phase pick up task node, a fully loaded truck after completion of the loading task must be returned to the dock. Constraints (5) - (8) are constraints on features with respect to OD. Specifically, constraints (5) and (6) ensure that the dock cannot be used to connect OD pairs. Constraints (7) and (8) impose a requirement that the starting and ending nodes of an OD pair cannot act as transit nodes connecting the OD pair.

To optimize driver and truck numbers, travel routes and schedules, in addition to driver binary route decision variables

And binary route decision variable for a truck>

We also need to define a binary decision variable ∈ whether the driver V ∈ V is traveling along the arc (i, j) in its trip, and whether the truck K ∈ K is traveling via the i ∈ N node to another node j ∈ N, respectively>

And a continuous time variable->

To indicate whether at least one truck crosses arc (i, j) and task node i e C service start time, respectively. Furthermore, with p _i ,i∈C ₁ Indicating the container unloading/loading time required by each customer. For the reader's convenience, the notations used in this study are presented in the appendix. The invention models the LCDP under IPOM as the following objective function:

where min represents the minimization function, x tableA binary route decision variable indicating driver, y a binary route decision variable of truck, a binary decision variable of truck crossing arc, β a binary decision variable of truck crossing on arc, s a time variable, λ ₁ Represents the daily fixed cost of the driver, V represents the set of drivers, j represents a task node, C represents the set of task nodes other than the dock,

binary route decision variables representing driver v on arc (0, j), 0 representing dock, λ ₂ Represents the daily fixed cost score of the truck, K represents the collection of trucks, and ` is `>

Representing a binary route decision variable, c 'for truck k on arc (i, j)' _ij Representing travel costs on arc (i, j) in lieu of a mode of transportation>

Represents the binary route decision variable for driver v on arc (i, j), Σ represents the sum, and e represents the belonging.

In particular, the first term in the objective function

The second term @, a cost of hiring the driver>

For truck deployment costs, the third term->

The fourth term ≥ takes into account the total fuel consumption cost of the truck for enqueuing fuel savings>

Travel costs for alternative modes of transportation.

The objective function is subject to constraints (1) - (8) and (10) - (29), the constraints (10) - (29) being as follows:

/>

wherein the content of the first and second substances,

representing a binary route decision variable for driver v on arc (0, i), N \ j represents a set of N minus j,

Binary route decision variable, C, representing truck k on arc (0, i) ₁ Is shown by D ₁ And P ₁ In the same client node, i' represents a task node in the same client node that corresponds to task node i, and +>

A binary decision variable indicating whether or not an arc (g, j) is traversed during a trip connecting an OD pair (o, d), g indicates a task node in C other than o and d, g ' indicates a task node in the same client node corresponding to task node g, C \ { g, g ' } indicates a set of C excluding g and g ', and>

indicates that the OD pair (o) is connectedD) whether or not an arc (o, i) is crossed on a trip, and>

a binary decision variable representing whether or not an arc (j, d) is crossed during a trip connecting an OD pair (o, d), and->

Binary decision variable, s, representing whether or not to cross arc (i, j) in the run connecting OD pairs (o, d) _i Representing the service start time, p, of the task node i _i Indicating the time of discharge/loading, s, of the task node i _i′ Represents the service start time, t, of the task node i _0i Representing the travel time, β, of the driver by truck transfer from dock 0 to the service node i _0i Binary decision variable, t ', representing the passing of a truck over arc (0,i)' _0i Representing the time, t, of the driver's journey through the alternate mode of transportation on arc (0, i) _i0 Represents the time the driver travels by truck over arc (i, 0), T represents the maximum work time per day, C ₂ Represents D ₂ And P ₂ Union of (1), s _j Represents the service start time, t, of task node j _ij Representing the time, t ', of the driver's trip through the truck on arc (i, j) ' _ij Representing the travel time of the driver on arc (i, j) through the alternate mode of transportation, and M represents a constant.

Where M is a sufficiently large number. The objective function shown in constraint (9) is the sum of the total operating costs, including driver employment costs, truck deployment costs, total truck fuel consumption costs to allow for queuing fuel savings, and travel costs resulting from alternative modes of transportation. The constraints (10) - (12) constitute a driver driving route. Specifically, the constraint (10) forces each driver to leave the dock at most once. Constraints (11) specify that each task node is visited once by one driver. Constraints (12) ensure that each driver conserves flow. The constraints (13) - (15) constitute a truck driving route. Constraints (13) limit each truck from being repeatedly deployed. The constraint (14) ensures that at least one truck remains at the customer site until the loading/unloading operation is completed. Constraints (15) ensure conservation of flow per truck. A constraint (16) limits the maximum trucking size. Constraints (17) - (19) specify the relationship between driver routes and truck movement routes between task nodes. Specifically, the constraint (17) specifies that the truck cannot be moved without driver guidance. Constraint (18) determines a binary decision variable beta _ij To indicate whether there is at least one truck crossing arc (i, j). The constraint (19) means that the driver can pass through the arc without a truck. Constraints (20) - (23) track truck flow. The constraint (20) is the conservation of flow for each OD pair. Constraints (21) ensure flow balance for each OD pair. Constraints (22) ensure that for each mission node, the number of arcs connecting the OD pair through that mission node is equal to the truck traffic. Constraint (23) indicates that the arc used to connect the OD pairs is traversed by at least one truck. Constraints (24) - (27) are constraints on task node scheduling. Constraints (24) ensure that the required loading/unloading time must be met. Constraints (25) and (26) dictate that each task must be completed within the driver's maximum daily working time. The constraint (27) specifies a service start time relationship between two customers that are consecutively visited by the same driver. Constraints (28) - (29) define the value ranges of the decision variables.

Step S300 specifically includes:

s310, constructing a character string for a driver, and sequentially inserting task nodes which are not served from a wharf; when a task node which is not yet served is inserted, a task node which is closest to the last inserted task node is selected to be inserted.

And S320, when the task node inserted in the character string of the driver reaches the maximum truck rank size constraint, ending the character string of the driver.

S330, constructing a character string for the next driver, and sequentially inserting the task nodes which are not served from the wharf until all the task nodes are inserted into the character string to obtain an initial solution corresponding to the network diagram.

In particular, the present invention contemplates a new mode of container transport operation that is more complex than the existing modes of operation. In particular, empty truck sharing between customers, truck queuing travel patterns, and allowing drivers to use other available modes of transportation besides trucks will make calculation of the actual number of drivers and trucks required, the manner in which trucks move between customers, the time drivers travel between customers, and the number of trucks (traveling together) on each arc more difficult. Therefore, a special heuristic construction algorithm (HCSA) is designed according to the problem characteristics to solve the proposed problem. The framework of the HCSA is shown in fig. 2. A greedy algorithm is utilized to construct the initial solution. Specifically, a string is constructed for each driver and an unserviceable task node will be selected based on the closest distance from the dock to the last intervening node. If any task node that is not yet in service cannot be inserted without violating the maximum truck bank size constraint, the current string is ended. This process will start again at the dock and continue to construct the next string until all task nodes have been inserted into the string. Maintaining the feasibility of truck bank size constraints during the generation of the initial solution is a challenge because the actual number of trucks occupied in each string (i.e., the fleet size of each string) will vary dynamically depending on the type of next task node attempting to be inserted into the string:

(1) For each task node i ∈ D ₁ The trucks required for this can only be met by the dock. Thus, fleet size has increased by 1.

(2) For each task node i ∈ D ₂ If i' e.d ₁ The vehicle fleet scale is kept unchanged if the vehicle fleet scale is contained in the character string; otherwise, fleet size would increase by 1.

(3) For each task node i ∈ P ₁ The demand may be satisfied by the dock or by an empty truck previously released from the second stage delivery task node. Thus, if the current string contains at least one available empty truck, the fleet size remains unchanged; otherwise, fleet size would increase by 1.

(4) For each task node i ∈ P ₂ If i' e.P ₁ If the character string is contained in the character string, the size of the motorcade is kept unchanged; otherwise, fleet size would increase by 1.

Step S400 specifically includes:

and S410, when the initial solution does not meet the constraint condition of task node scheduling, re-determining the initial solution corresponding to the network graph.

And S420, when the initial solution meets the constraint condition of task node time arrangement, evaluating the initial solution based on the objective function to obtain an evaluation result.

And S430, when the evaluation result meets a preset requirement, taking the initial solution as an updating solution corresponding to the network diagram.

Specifically, it is determined whether the constraint condition of the truck bank size is met when the initial solution is determined, and since the unmanned truck is driven under the lead of the lead truck, but the number of trucks in the truck queue is constrained, the requirement of being less than or equal to the maximum number L of trucks in the truck queue needs to be met. When this requirement is met, an initial solution is generated. And then carrying out practical feasibility test, wherein the constraint conditions (24) - (27) are the constraint conditions of task node scheduling, and if the initial solution does not meet the constraint conditions of task node scheduling, returning to the step S300 and re-determining the initial solution. If the initial solution meets the constraints of the task node scheduling, the initial solution is evaluated based on the objective function to obtain an evaluation result, for example, driver hiring cost, truck deployment cost, total truck fuel consumption cost considering the queuing oil saving, and travel cost generated by the alternative transportation mode can be evaluated.

For example, the total operation cost is used for evaluation, the preset requirement may be that the total operation cost is less than the preset cost, and when the evaluation result meets the preset requirement, it indicates that the quality of the initial solution is high, and the initial solution may be used as an update solution.

Although the heuristic solving method can obtain a feasible solution, the heuristic solving method may not obtain a high-quality solution, so that a simulated annealing mechanism is integrated into the heuristic solving algorithm to diversify the structure of the solution and guide to search for a better solution in an effective iteration mode.

Step S400 further includes:

s440, when the evaluation result does not meet the preset requirement, searching the initial solution by adopting a neighborhood search operator to obtain a neighborhood solution; wherein the neighborhood search operator comprises: an exchange operator, an inversion operator, and an insertion operator.

S450, when the objective function value corresponding to the domain solution is smaller than the objective function value corresponding to the initial solution, taking the neighborhood solution as the initial solution corresponding to the network map, and reducing the temperature corresponding to the initial solution.

And S460, when the initial solution meets the termination condition, taking the initial solution as an updating solution corresponding to the network graph.

Specifically, when the evaluation result does not meet the preset requirement, the initial solution needs to be updated. Three different neighborhood search operators are used, namely swap, invert, and insert. In particular, the swap operator is the position of two randomly selected nodes in a string of characters in the swap solution. The reversal is performed by randomly selecting a pair of nodes in the string and then reversing the order of the substrings between them. The insertion operator removes the randomly selected node and reinserts it into another location in the same string. Simulated annealing will randomly select one from the three neighborhood search operators in each iteration to obtain a better domain solution.

Step S400 further includes:

s470, when the objective function value corresponding to the domain solution is greater than or equal to the objective function value corresponding to the initial solution, determining the initial solution corresponding to the network map according to Metropolis criterion, and reducing the temperature corresponding to the initial solution.

And S480, when the initial solution does not meet the termination condition, continuously judging whether the initial solution meets the constraint condition of task node time arrangement or not, and when the initial solution meets the evaluation result and meets the preset requirement or the initial solution meets the termination condition, taking the initial solution as an updating solution corresponding to the network graph.

Specifically, the termination condition includes: and the temperature corresponding to the initial solution is smaller than a preset temperature threshold value or the iteration number of the initial solution reaches the maximum continuous non-improvement iteration number. Decreasing the temperature corresponding to the initial solution according to the annealing rateAnd (4) low temperature. Simulated annealing algorithm (SA) starts from an initial solution of a heuristic construction S ⁰ . Then, the current solution S (from the initial solution S) is passed through a predefined neighborhood structure ⁰ Start) is iteratively converted into its neighborhood solution S'. If the objective function value of S 'is less than that of S, we will replace the current solution S with the domain solution S'; otherwise, we will accept a new solution with Metropolis probability, e.g.,

where Δ H represents the difference between the objective function values for solution S 'and solution S (e.g., obj (S') -obj (S)), and T is the current temperature. From the above, it can be seen that the SA algorithm allows accepting bad solutions with less probability, which effectively reduces the possibility of trapping in local minima. This search process is repeated until a stop condition, i.e. a pre-specified temperature threshold T, is met ^end Or maximum number of consecutive non-increasing iterations B _max . Finally, the optimal solution S is returned ^* . The essence of the SA algorithm is to bring the temperature from a given initial temperature T ⁰ Gradually decreases to a specified termination temperature T ^end . After a number of successive attempts, the current temperature T will be lowered according to the cooling rule T ← T · epsilon, where epsilon is the annealing rate, i.e. a constant between 0 and 1, to control the cooling rate, in order to obtain a better solution.

Detailed description of the preferred embodiment

We performed a computational experiment to evaluate the effectiveness of the proposed model and algorithm. The experiment used randomly generated examples to test the model and algorithm. The customer randomly selects within a square area of 200km in length, with the wharf centered in the area and having coordinates of (100 ). Each customer has only one delivery or pick-up request, where the required time to unload or load, in hours, is randomly generated from the set 2,3, \ 8230;, 5. This chapter designates motorcycles as an alternative means of transportation for the driver to use at any customer location. Considering that the running speed of a truck is 60km/h and the running speed of a motorcycle is 35km/h, the running time between two customers can be obtained according to the euclidean distance between the two customers. According to consideration of truckExisting parameter settings for queued LCDP, herein set model parameters as follows: the fuel saving coefficient η, the maximum number of trucks in line L and the maximum driver's daily operating time T forming a bank are set to 0.1, 6 and 16, respectively. As for the cost-related parameter, the daily fixed cost λ of hiring a driver and a truck ₁ And λ ₂ Set to 100 and 50 respectively. Cost per fuel consumption of truck and per travel cost of motorcycle c _ij And c' _ij Set to 1 and 0.5, respectively. Based on some preliminary experiments and SA criteria, we set the following algorithm parameters: initial temperature T ⁰ Temperature threshold T ^end Annealing rate ε and maximum number of consecutive non-increasing iterations B _max And set to 20, 0.0001, 0.9999 and 3000, respectively, in the experiment. Furthermore, T is used ^end =0.0001 and B _max =3000 as a stop condition of HCSA. Based on these parameter settings, we performed numerical experiments on sets of random examples with customer sizes varying from 4 to 400. Each set of examples is named "a + b," where a and b represent the number of delivery and pickup customers, respectively. For each customer scale, five random examples will be generated and the average solution results will be presented.

To evaluate the solution performance of using HCSA to solve LCDP-IPOM, we will compare the results of using HCSA and solving the proposed model directly using the CPLEX commercial solver. The maximum run time for CPLEX to solve each of the examples was set to 1 hour. Note that the HCSA algorithm was solved 3 times for each example and the average calculation was shown. Table 2 gives the results of the HCSA algorithm and CPLEX solution, with the minimum value for each index highlighted in bold. For both methods, we show the objective function value (Obj) obtained within the time limit and the elapsed CPU run time (time). We will also show the mean target value Gap (Gap) between HCSA and CPLEX. As shown in table 2, we found that CPLEX could not find a feasible solution for over 30 customer's calculations in 1 hour, and only the smallest scale example could be solved directly by CPLEX. In contrast, the proposed HCSA average can find a high quality solution for all the examples within 8 minutes. More specifically, for most cases where CPLEX can be solved within time constraints, the HCSA can achieve the same or better solution, i.e., average 696 v.s.3004s, in a much shorter time than CPLEX. For the CPLEX difficult to solve example, HCSA found a high quality solution only within 0.3 hours on average. In particular, HCSA may achieve a good solution for the calculation scale up to 400 (200 + 200) in 43 minutes. The above findings demonstrate on the one hand the considerable computational complexity of the proposed problem and the effectiveness and efficiency of the proposed algorithm in solving the problem. To further illustrate the computational performance of HCSA, we visualize the variation of CPU run time of HCSA versus customer size in fig. 3. One notable observation is that CPU time grows linearly as customer size increases. This is consistent with our expectation that the proposed heuristic based on simulated annealing takes advantage of meta-heuristics and can find reasonably good solutions in a reasonable time, which suggests that HCSA has great potential in the practical application of LCDP.

Table 2 hcsa and CPLEX solution result comparison

/>

Note that Gap = (Obj) ¹ -Obj ² )/Obj ² ×100％

Based on any embodiment of the novel local container transportation typesetting optimization method in the truck queuing mode, the invention further provides an embodiment of computer equipment.

The computer device of the present invention includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

acquiring a network map; wherein the network graph comprises a set of nodes and a set of arcs, wherein the nodes in the set of nodes comprise a delivery client node, a pickup client node and a dock, and the arcs in the set of arcs are connecting arcs between two of the nodes; the delivery client node comprises: the task nodes before unloading and the task nodes after unloading, wherein the goods-taking client node comprises: a task node before loading and a task node after loading;

determining an initial solution corresponding to the network diagram;

Based on any one of the embodiments, the invention further provides an embodiment of a computer-readable storage medium.

The computer-readable storage medium of the present invention, on which a computer program is stored, which, when executed by a processor, implements the steps of:

acquiring a network map; wherein the network graph comprises a set of nodes and a set of arcs, the nodes in the set of nodes comprise a delivery client node, a pickup client node and a dock, and the arcs in the set of arcs are connecting arcs between two nodes; the delivery client node comprises: the task nodes before unloading and the task nodes after unloading, wherein the goods-taking client node comprises: a task node before loading and a task node after loading;

determining an objective function of the total operation cost according to the network graph; wherein the total operating cost comprises: driver hiring costs, truck deployment costs, total fuel consumption costs for trucks to allow for queuing fuel savings, and travel costs resulting from alternative modes of transportation;

determining an initial solution corresponding to the network diagram;

It is to be understood that the invention is not limited to the examples described above, but that modifications and variations may be effected thereto by those of ordinary skill in the art in light of the foregoing description, and that all such modifications and variations are intended to be within the scope of the invention as defined by the appended claims.

Claims

1. A method for optimizing transportation and layout of a local container in a novel truck queuing mode is characterized by comprising the following steps:

acquiring a network map; the network graph comprises a node set and a set of arcs, wherein the nodes in the node set comprise delivery client nodes, pickup client nodes and wharfs, and the arcs in the set of arcs are connecting arcs between any two nodes in the node set; the delivery client node comprises: the task nodes before unloading and the task nodes after unloading, wherein the goods-taking client node comprises: the task nodes before loading and the task nodes after loading;

determining an initial solution corresponding to the network diagram;

under the constraint condition, updating the initial solution to minimize the objective function to obtain an updated solution corresponding to the network diagram;

the objective function is:

where min represents a minimization function, x represents a binary route decision variable for the driver, y represents a binary route decision variable for the truck, α represents a binary decision variable for the truck to traverse the arc, and β represents the presence of an arc on the arcBinary decision variable for truck passing, s represents time variable, λ ₁ Represents the driver's daily fixed cost, V represents the driver, V represents the set of drivers, j represents a task node, C represents the set of task nodes other than the dock,

binary route decision variable representing driver v on arc (0, j), 0 representing dock, λ ₂ Represents the daily fixed cost score of the truck, K represents the collection of trucks, and ` is `>

Binary route decision variables, c, representing truck k on arc (0, j) _ij Represents the fuel consumption cost of the truck on arc (i, j), i represents the mission node, N represents the set of nodes comprising the dock, β _ij Binary decision variables indicating that there is a truck crossing on arc (i, j), η indicates the fuel savings rate for trucks in the truck fleet column, and &>

Representing a binary route decision variable, c 'for truck k on arc (i, j)' _jj Represents the travel cost on arc (i, j) instead of the mode of transportation, based on the number of hours on the trip>

2. The optimization method for local container transportation typesetting in the novel truck queue driving mode according to claim 1, wherein the constraint conditions are as follows:

/>

wherein the content of the first and second substances,

a binary decision variable indicating whether or not an arc (i, D) is crossed in a trip connecting OD pairs (0, D), D indicating a task node of a delivery task, D ₁ Represents a set of task nodes before a dump, and->

A binary decision variable indicating whether or not an arc (0, i) is crossed in a trip connecting an OD pair (o, d), o indicating the node where the task starts, R indicating the set of OD pairs, </or >>

Binary decision variable representing whether or not an arc (i, 0) is crossed in a trip connecting an OD pair (o, d), based on a predetermined criterion>

A binary decision variable indicating whether or not to cross an arc (i, o) on a trip connecting an OD pair (o, d), N \ o indicating the set of N minus o, and/or>

Represents the travel of the connected OD pair (0, d)A binary decision variable of whether or not to cross arc (d, i), N \ d representing the set of N minus d, and/or->

Indicates presence, or is present>

Binary route decision variable, C, representing truck k on arc (0, i) ₁ Represents D ₁ And P ₁ Represents a task node corresponding to task node i in the same client node, i' is greater than or equal to>

Represents a binary route decision variable of a truck k on an arc (j, i), N \ { j, j ' } represents a set of N excluding j and j ', j ' represents a task node corresponding to the task node j in the same client node, L represents the maximum number of trucks in a truck queue, and>

binary decision variable representing whether or not an arc (i, g) is crossed in a trip connecting an OD pair (o, d), based on the number of consecutive decision variables, and based on the number of consecutive decision variables>

A binary decision variable indicating whether or not an arc (g, j) is crossed in a trip connecting an OD pair (o, d), g indicates a task node other than o and d in C, g 'indicates a task node corresponding to task node g in the same client node, C \ g, g' } indicates a set of C excluding g and g ',' and>

binary decision variable representing whether or not an arc (o, i) is crossed in a trip connecting an OD pair (o, d), based on a predetermined criterion>

Binary decision variable, s, representing whether or not to cross arc (i, j) in the run connecting OD pairs (o, d) _i Representing the service start time, p, of the task node i _i Indicating the time of discharge/loading, s, of the task node i _i′ Representing the service start time, t, of the task node i _0i Representing the travel time, β, of the driver by truck transfer from dock 0 to the service node i _0i Binary decision variable, t ', representing the passing of a truck over arc (0,i)' _0i Representing the travel time, t, of the driver on the arc (0, i) by an alternative mode of transportation _i0 Represents the time the driver travels by truck over arc (i, 0), T represents the maximum work time per day, C ₂ Is shown by D ₂ And P ₂ Union of (1), s _j Represents the service start time, t, of task node j _ij Representing the time, t 'the driver travelled by the truck over arc (i, j)' _ij Represents the travel time of the driver through the alternate mode of transportation on arc (i, j), and M represents a constant.

3. The method for optimizing local container transportation composition in the novel truck queue driving mode according to any one of claims 1-2, wherein the determining the initial solution corresponding to the network diagram comprises:

when the task node inserted in the driver character string reaches the maximum truck number constraint in the truck queue, ending the driver character string;

4. The optimization method for the layout of local container transportation in the new truck queuing mode according to any one of claims 1-2, wherein the updating the initial solution to minimize the objective function under the constraint condition to obtain the updated solution corresponding to the network diagram includes:

when the initial solution does not meet the constraint condition of task node scheduling, re-determining the initial solution corresponding to the network graph;

5. The method of optimizing local container transportation composition in the new truck queuing mode according to claim 4, wherein the updating the initial solution to minimize the objective function under constraint conditions to obtain the updated solution corresponding to the network map further comprises:

when the objective function value corresponding to the neighborhood solution is smaller than the objective function value corresponding to the initial solution, taking the neighborhood solution as the initial solution corresponding to the network graph, and reducing the temperature corresponding to the initial solution;

6. The method of optimizing local container transportation composition in the new truck queuing mode according to claim 5, wherein the updating the initial solution to minimize the objective function under constraint conditions to obtain the updated solution corresponding to the network map further comprises:

when the objective function value corresponding to the neighborhood solution is larger than or equal to the objective function value corresponding to the initial solution, determining the initial solution corresponding to the network graph according to the Metropolis criterion, and reducing the temperature corresponding to the initial solution;

and when the initial solution does not meet the termination condition, continuously judging whether the initial solution meets the constraint condition of task node time arrangement or not, and when the initial solution meets the evaluation result and meets the preset requirement or the initial solution meets the termination condition, taking the initial solution as an updating solution corresponding to the network graph.

7. The method of optimizing local container transport composition in truck queue mode according to claim 6, wherein the termination condition comprises: the temperature corresponding to the initial solution is smaller than a preset temperature threshold value or the iteration number of the initial solution reaches the maximum continuous non-improvement iteration number;

8. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.

9. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.