Multi-target evaluation system optimization method for sorting small pieces in logistics transfer field

Abstract

The invention discloses a multi-target evaluation system optimization method for sorting small pieces in a logistics transfer field. The method comprises the following steps: designing and defining multi-dimensional indexes of a sorting plan # imgabs0 # for mapping and distributing grid resources and package plate resources; performing normalization processing on the multi-dimensional indexes, and establishing a multi-target sorting performance evaluation system; through an industrial internet platform and a digital twin system, standard acquisition and real-time transmission of actual sorting data are realized; the mapping between the package flow direction and the grids is optimized, and the optimization part is divided into two stages: the first stage is global number distribution optimization; and in the second stage, the number distribution in the region is optimized. According to the method, on the basis of establishment of a multi-target evaluation system, global optimization is guaranteed and local search precision is improved on the basis of a hybrid algorithm framework of a genetic algorithm and a simulated annealing algorithm, so that a complex resource allocation problem in logistics sorting is effectively solved.

Description

Technical Field

[0001] The present invention relates to the field of logistics sorting, and in particular to an optimization method for a multi-objective evaluation system for sorting small items in a logistics transfer yard. Background Art

[0002] With the rapid development of e-commerce and express logistics in recent years, small-item sorting operations in logistics transfer yards face a series of problems such as high sorting volume, sorting errors, uneven utilization of sorting resources, and reflux processing. In the existing technology, the evaluation indicators for the sorting system are often single, failing to fully reflect the overall operation status of the sorting system from multiple angles such as sorting capacity, sorting accuracy, reflux rate, and sorting grid utilization rate, and there is a lack of effective connection between the evaluation results and resource allocation. At the same time, the development of digital twin technology has made it possible to monitor and predict the logistics sorting process in real time. How to build a multi-objective evaluation system and use meta-heuristic optimization algorithms (such as genetic algorithms and simulated annealing algorithms) to dynamically and accurately allocate sorting resources has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a method for optimizing a multi-objective evaluation system for small-piece sorting in a logistics transfer station. Based on the establishment of a multi-objective evaluation system, a hybrid algorithm framework based on a genetic algorithm and a simulated annealing algorithm is established to ensure global optimization and improve local search accuracy, thereby effectively solving complex resource allocation problems in logistics sorting.

[0004] The object of the present invention is achieved through the following technical solution: A method for optimizing a multi-objective evaluation system for sorting small items in a logistics transfer station, comprising the following steps: Sorting plan for mapping and allocating slot resources and package resources Design and define multidimensional indicators; Normalize the multi-dimensional indicators and establish a multi-objective sorting performance evaluation system; Through the industrial Internet platform and digital twin system, the actual sorting data can be standardized and transmitted in real time; The mapping between package flow and grid openings is optimized, and the optimization part is divided into two stages: the first stage is the global quantity distribution optimization; the second stage is the internal quantity distribution optimization of the region.

[0005] The beneficial effects of the present invention are as follows: in view of the fact that the traditional single sorting system evaluation index is difficult to fully reflect the problem in the sorting process, the present invention constructs a multi-objective sorting performance evaluation system, and through strict mathematical definition, normalization and weight allocation of various indicators, the overall evaluation function can fully and objectively reflect the operation effect of the sorting system. In addition, the digital twin technology is used to build a virtual model of the logistics transfer site and the sorting process, and the data of each key indicator is obtained in real time through the Internet of Things sensors and data acquisition equipment to ensure that the data based on the evaluation system is accurate and real-time. Based on the established multi-objective evaluation system, a hybrid algorithm framework based on genetic algorithm and simulated annealing algorithm is proposed to ensure global optimization and improve local search accuracy, thereby effectively solving complex resource allocation problems in logistics sorting. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0007] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0008] In recent years, with the rapid development of e-commerce and express logistics, the volume of parcels in the logistics network has continued to increase rapidly. As a hub and centralized processing station in the logistics network, the logistics transfer yard plays a vital role. It is not only the core node for the centralized distribution, allocation and re-dispatch of parcels, but also undertakes the important tasks of improving global logistics efficiency, reducing transportation costs and ensuring delivery timeliness. In the transfer yard, a large number of parcels are gathered and diverted in the shortest time, and preliminarily sorted according to the destination, providing strong support for subsequent rapid delivery. Logistics sorting, as an important link in the operation of the transfer yard, plays a key role in connecting the upper and lower parts of the entire logistics process. By adopting advanced cross-belt sorting machines and intelligent systems, accurate division and efficient diversion of parcels can be achieved, which can not only significantly improve the sorting speed and accuracy, but also effectively reduce the problems of return flow and resource waste caused by misclassification or repeated sorting. Specifically: like Figure 1 As shown, a multi-objective evaluation system optimization method for sorting small items in a logistics transfer station includes the following steps: First, we will map and allocate the current grid resources and package resources to the sorting plan based on the core key indicators of logistics sorting: sorting capacity indicator, sorting accuracy indicator, sorting return rate indicator, and sorting grid utilization indicator. Design and definition of multidimensional indicators: 1. Sorting capacity: Sorting capacity indicators include instant sorting capacity, peak sorting capacity, and steady-state sorting capacity. The instant sorting capacity is the number of packages sorted per minute under the sliding window, which is defined as: in, T is the length of the sliding window in minutes, For in time The number of parcels sorted within minutes; Indicates the end time of the sorting shift.

[0009] Peak sorting capacity is defined as the highest capacity value that occurs during the entire sorting shift, defined as Steady-state capacity is defined as the average capacity during the steady-state period of the sorting system, where the system steady-state period is defined as the time period when the instantaneous capacity is between 80% and 100% of the peak capacity. The specific definition is as follows: 2. Sorting accuracy: Sorting accuracy is defined as the ratio of the number of parcels correctly sorted into the designated slot to the total number of parcels sorted during the sorting shift. 100% in, is the total number of parcels sorted in this sorting shift, The number of parcels correctly sorted into the designated slot during the sorting shift.

[0010] 3. Sorting return rate: The sorting return rate is defined as the number of packages that are scanned more than the upper limit on the conveyor belt during the sorting process. Returns are usually caused by uneven allocation of packages and slot resources, so this indicator is required to be reduced as much as possible. It is defined as: 100% in, It is the total number of packages returned in this shift.

[0011] 4. Sorting grid utilization rate: Considering that the personnel in the site are evenly distributed in different sorting areas, the workload of the personnel in each area needs to be as even as possible, that is, the number of grid locks in each area should be as equal as possible in the current shift. J Sorting areas, each area j In For the area j Middle i The number of times a slot is locked in the current shift is = 1,2,..., . On this basis, for the region j The average number of locks for all cells is . Therefore, when the grid utilization rate is the most balanced, the area j The ideal number of locks for all the compartments in the In order to characterize the area j The actual degree of discreteness of sorting within the region is determined by j The standard deviation of the number of locks in the grid is used to model: Among them, the more balanced the number of times all the grids are locked, The closer it is to 0. Therefore, we use the idea of ​​coefficient variation to construct a regional uniformity index , defined as: On this basis, considering the uniform utilization of sorting slots in each area of ​​the entire logistics transfer station, we use the average uniformity index in each area to characterize the uniform utilization of the overall sorting slots, which is defined as: Normalize the multi-dimensional indicators and establish a multi-objective sorting performance evaluation system; Through these four indicators, we can comprehensively evaluate the sorting performance of the logistics sorting system from the perspectives of efficiency, quality, stability and resource utilization in the logistics transfer station. Based on the comprehensive definition of the above four indicators, we normalize the above indicators to form an evaluation vector: Among them, the sorting accuracy, sorting return rate, and slot utilization rate indicators are normalized indicators, so only the capacity indicator needs to be normalized: in, as well as In order to obtain the theoretical peak capacity and the upper bound of the steady-state capacity by inputting the same real parcel information (specific arrival of parcels, parcel flow distribution, etc.) under the sorting shift with the help of the established digital twin system, the lock time is set to 0. On this basis, we set the weight vector to , then the multi-objective sorting performance evaluation system can be expressed as: Or it can be directly expressed as: in, When k=1,2,3,4, it means Through the industrial Internet platform and digital twin system, the actual sorting data can be standardized and transmitted in real time; Based on the established multi-objective sorting performance evaluation system, the industrial Internet platform and digital twin system are used to realize the standardized collection and real-time transmission of actual sorting data. Specifically, the sorting data is collected in real time through sensors in the cross-belt sorter, handheld RFID devices in the logistics transfer yard, and IoT devices such as cameras in the yard, and the number of sorted packages per unit time, the accurate number of package sorting, the number of reflux, and the number of packages dropped in each grid in the sorting area are dynamically obtained, thereby providing real-time and accurate data support for the multi-objective evaluation system.

[0012] On this basis, we propose a hybrid algorithm framework based on genetic algorithm and simulated annealing algorithm to optimize the mapping between package flow and sorting slots. The optimization part is divided into two stages: 1. Optimization of the overall quantity between the overall package flow and the sorting slots (global quantity distribution optimization) 2. When the global quantity is determined, the quantity between the package flow and the sorting slots in each sorting area is optimized (internal quantity distribution optimization in the area). The overall algorithm idea is as follows: Optimize the overall quantity between the overall package flow and the sorting grid 1. Considering the stability of the parcel flow between sorting shifts, we use the mapping of the overall parcel flow in the current most recent sorting shift and the overall number of sorting slots as the initial population and the initial population as the candidate solution. x , candidate solution x is a vector that represents the number of slots allocated to each package flow and satisfies in, The total number of sorting slots.

[0013] 2. For candidate solutions x , the sorting plan for the candidate solution is solved through the digital twin system (Mapping relationship between package flow and grid opening) Conduct multi-objective evaluation system evaluation and then calculate the overall comprehensive index Then the fitness in the genetic algorithm is defined as F ( x ) = Q ( x ).

[0014] 3. Genetic algorithm global search: candidate solutions in the population are searched according to fitness F ( x ) to perform selection, crossover and mutation to generate a new generation of population. In order to ensure the constraint (the total number of packages remains unchanged), the crossover and mutation operations need to be designed or coordinated with the repair operation.

[0015] 4. Simulated annealing local search: For some elite candidate solutions, perform local search in their neighborhood. The specific operations are: a. Select the parcel flow directions that contribute too much or too little to the objective function from the current candidate solutions; b. Generate candidate solutions by transferring a grid from a parcel flow direction that contributes too much to the objective function to a parcel flow direction that contributes too little ; c. Calculate the difference if , then accept; if , then with probability exp( )accept; d. Lower the temperature The above process is repeated until the temperature is lower than the set lower limit.

[0016] 5. Repeat genetic operations and local annealing optimization, and continuously update the population until the convergence conditions (such as maximum number of generations or insufficient fitness improvement) are met. The candidate solution at this time is the global solution. .

[0017] Optimization of quantity distribution within the region 1. In the obtained global solution In the sorting grid, each sorting grid is divided according to the sorting area to which it belongs. j , among which the candidate solutions The total number of occupied slots is fixed. Considering that the sorting capacity index cannot be quantitatively estimated within the region, the multi-objective evaluation system in each region is adjusted to 2. In each sorting area j In the local population (the initial perturbation solution is derived from the global solution), a genetic algorithm (selection, crossover, mutation) is used to generate multiple candidate allocation schemes in the region, and the corresponding .

[0018] 3. Similar to the global search, the simulated annealing algorithm is used to perform local search adjustments on the elite candidate solutions in the region.

[0019] 4. Take each sorting area j The best solution finally obtained , updated to the global solution In this way, we can obtain the global and detailed optimization solutions that specifically meet the multi-objective evaluation system.

[0020] In summary, the present invention constructs an intelligent logistics optimization architecture based on a multi-objective sorting performance evaluation system and a digital twin platform, proposes a hybrid genetic and simulated annealing meta-heuristic algorithm framework, and realizes dynamic collaborative optimization among key performance indicators in the sorting system, providing a new theoretical basis and technical method for the comprehensive scheduling and intelligent decision-making of logistics sorting systems.

Claims

1. A multi-objective evaluation system optimization method for sorting small items in a logistics transfer station, characterized by: The following steps are involved: Sorting plan for mapping and allocating slot resources and package resources Design and define multidimensional indicators; Normalize the multi-dimensional indicators and establish a multi-objective sorting performance evaluation system; Through the industrial Internet platform and digital twin system, the actual sorting data can be collected in a standardized manner and transmitted in real time; The mapping between package flow and grid openings is optimized, and the optimization part is divided into two stages: the first stage is the global quantity distribution optimization; the second stage is the internal quantity distribution optimization of the region.

2. According to claim 1, a multi-objective evaluation system optimization method for sorting small items in a logistics transfer station is characterized by: The multi-dimensional indicators adopt core key indicators for logistics sorting, including sorting capacity indicators, sorting accuracy indicators, sorting return rate indicators, and sorting grid utilization indicators.

3. The optimization method of a multi-objective evaluation system for sorting small items in a logistics transfer station according to claim 2 is characterized by: The sorting plan for mapping and allocating the grid resources and the package resources The multidimensional indicators designed and defined include: (1) Sorting capacity indicators, including instant sorting capacity, peak sorting capacity, and steady-state sorting capacity; the instant sorting capacity is the number of packages sorted per minute under the sliding window, defined as: in, T is the length of the sliding window in minutes, For in time The number of parcels sorted within minutes, Indicates the end time of the sorting shift; Peak sorting capacity is defined as the highest capacity value that occurs during the entire sorting shift, defined as: Steady-state capacity is defined as the average capacity during the steady-state period of the sorting system, where the system steady-state period is defined as the time period when the instantaneous capacity is between 80% and 100% of the peak capacity, defined as: (2) The sorting accuracy index is defined as the ratio of the number of parcels correctly sorted into the designated slot to the total number of parcels sorted during the sorting shift: 100% in, is the total number of parcels sorted in this sorting shift, The number of parcels correctly sorted into the designated slots during the sorting shift; (3) The sorting return rate index is defined as the number of packages scanned on the conveyor belt that exceeds the upper limit during the sorting process. Returns are usually caused by uneven allocation of packages and slot resources, so this index is required to be reduced as much as possible and is defined as: 100% in, The total number of parcels returned during this shift; (4) Sorting slot utilization rate: Assume that the current logistics transfer station has J Sorting areas, each area j In For the area j Middle i The number of times a slot is locked in the current shift is: = 1,2,..., . For Region j The average number of locks for all the cells is: In order to depict the area j The actual degree of discreteness of sorting within the area, using the j The standard deviation of the number of locks in the grid is used to model: Among them, the more balanced the number of times all the grids are locked, The closer it is to 0; using the idea of ​​coefficient variation, construct a regional uniformity index , defined as: The average uniformity index in each area is used to characterize the uniformity of the overall sorting grid utilization, which is defined as: 。 4. The optimization method of a multi-objective evaluation system for sorting small items in a logistics transfer station according to claim 3 is characterized by: The multi-dimensional indicators are normalized and the multi-objective sorting performance evaluation system is established, including: After normalization, the above indicators are formed into an evaluation vector: Among them, the sorting accuracy, sorting return rate, and slot utilization rate indicators are normalized indicators. Only the capacity indicator needs to be normalized: in, as well as The theoretical peak capacity and the upper bound of the steady-state capacity are obtained by inputting the same real parcel letter of the sorting shift and setting the lock time to 0 with the help of the established digital twin system; Set the weight vector to , then the multi-objective sorting performance evaluation system is expressed as: Or directly use the function expression as: in, When k=1,2,3,4, it means .

5. The optimization method of a multi-objective evaluation system for sorting small items in a logistics transfer station according to claim 3 is characterized by: The industrial Internet platform and digital twin system are used to realize standardized collection and real-time transmission of actual sorting data, including: Sorting data is collected in real time through sensors in the cross-belt sorter, handheld RFID devices in the logistics transfer yard, cameras in the yard and other IoT devices, so as to dynamically obtain the number of parcels sorted per unit time, the accurate number of parcels sorted, the number of refluxes, and the number of parcels dropped into each grid in the sorting area.

6. The optimization method of a multi-objective evaluation system for sorting small items in a logistics transfer station according to claim 1 is characterized by: The global quantity distribution optimization, i.e., the overall quantity optimization between the overall package flow and the sorting grid, includes: A1. Considering the stability of the parcel flow between sorting shifts, the mapping of the overall parcel flow in the current most recent sorting shift and the overall number of sorting slots is used as the initial population, and the initial population is used as the candidate solution. x , candidate solution x is a vector that represents the number of slots allocated to each package flow and satisfies in, is the total number of sorting slots; i Indicates i Package flow, Indicates the number of slots allocated to the i-th package flow, A collection representing the direction of package flow; A2. For candidate solutions x , the sorting plan for the candidate solution is solved through the digital twin system Conduct multi-objective evaluation system assessment and then calculate the overall comprehensive index Then the fitness in the genetic algorithm is defined as F ( x ) = Q ( x ); A3. Genetic algorithm global search: Candidate solutions in the population are searched according to their fitness F ( x ) to perform selection, crossover and mutation to generate a new generation of population; in order to ensure the constraints, the total number of packages remains unchanged; A4. Simulated annealing local search: For some elite candidate solutions, local search is performed in their neighborhood. The specific operations are as follows: Select the parcel flow directions that contribute excessively or too little to the objective function from the current candidate solutions; Among them, the improvement value of the overall comprehensive index brought by the additional allocation of one grid to each package flow direction in the candidate solution is defined as the contribution value. Let the candidate solution ,in represents the number of slots allocated to the ith package flow direction, G is the total number of package flows; first define the standard basis vector as , only the i-th component of this vector is 1, then the candidate solution is x The contribution to the i-th component is ; Excessive contribution to the objective function means that the contribution value of the parcel flow direction is higher than the preset excess contribution threshold; under-contribution to the objective function means that the contribution value of the parcel flow direction is lower than the preset under-contribution threshold; b. Generate candidate solutions by transferring a grid from a parcel flow direction that contributes too much to the objective function to a parcel flow direction that contributes too little ; c. Calculate the difference if , then accept; if , then with probability exp( )accept; d. Lower the temperature Repeat the above process until the temperature is lower than the set lower limit; A5. Repeat genetic operations and local annealing optimization, and continuously update the population until the convergence condition is met. The candidate solution at this time is the global solution. .

7. The optimization method of a multi-objective evaluation system for sorting small items in a logistics transfer station according to claim 6 is characterized by: The optimization of the quantity distribution within the region, i.e., the optimization of the quantity between the parcel flow direction and the sorting grid in each sorting region when the global quantity is determined, includes: B1. In the obtained global solution In the process, each sorting grid is divided according to the sorting area to which it belongs; for each sorting area j , among which the candidate solutions The total number of occupied slots is fixed; considering that the sorting capacity index cannot be quantitatively estimated within the region, the multi-objective evaluation system in each region is adjusted to: B2. In each sorting area j Generate local species, that is, derive the initial perturbation solution from the global solution, use genetic algorithm to generate multiple candidate allocation schemes in the region, and calculate the corresponding ; B3, use simulated annealing algorithm to perform local search adjustment on the elite candidate solutions in the region; B4 takes each sorting area j The best solution finally obtained , update to the global solution In this way, we can obtain the global and detailed optimization solutions that specifically meet the multi-objective evaluation system.

Images