CN114493047B - Many-to-many goods scheduling optimization distribution method based on available resource coding - Google Patents

Many-to-many goods scheduling optimization distribution method based on available resource coding Download PDF

Info

Publication number
CN114493047B
CN114493047B CN202210352848.5A CN202210352848A CN114493047B CN 114493047 B CN114493047 B CN 114493047B CN 202210352848 A CN202210352848 A CN 202210352848A CN 114493047 B CN114493047 B CN 114493047B
Authority
CN
China
Prior art keywords
goods
gene
many
coding
warehouse
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202210352848.5A
Other languages
Chinese (zh)
Other versions
CN114493047A (en
Inventor
赵国宏
高润芳
江光德
赵云飞
芮铁军
姜伟
魏建光
冯国宝
宫树香
高军强
梁辉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Pla 96901
Original Assignee
Pla 96901
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Pla 96901 filed Critical Pla 96901
Priority to CN202210352848.5A priority Critical patent/CN114493047B/en
Publication of CN114493047A publication Critical patent/CN114493047A/en
Application granted granted Critical
Publication of CN114493047B publication Critical patent/CN114493047B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06312—Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/083—Shipping
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/087—Inventory or stock management, e.g. order filling, procurement or balancing against orders

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Theoretical Computer Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Development Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Evolutionary Biology (AREA)
  • Game Theory and Decision Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Biomedical Technology (AREA)
  • Physiology (AREA)
  • Genetics & Genomics (AREA)
  • Artificial Intelligence (AREA)
  • Educational Administration (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention belongs to the technical field of cargo scheduling, and particularly relates to a many-to-many cargo scheduling optimization distribution method. The technical scheme is as follows: when the optimal distribution of many-to-many goods scheduling is solved by utilizing a genetic algorithm, the method comprises a many-to-many goods scheduling optimal distribution method based on available resource coding, wherein the goods type is taken as a gene fragment and the available quantity is taken as a gene locus during coding, and a chromosome gene string form is determined; taking a non-zero value part in a road passable constraint sparse matrix as a set, recording the row and column index relation between warehouse goods and a market, encoding the row and column index relation according to positive integers, and determining a symbol encoding subset of a gene locus value; the cross operation is carried out among the same type of warehouse goods; mutation operations are performed in alternative coding subsets of gene segments. The invention solves the problem that two dimensions of distribution relation and distribution quantity are simultaneously used for gene coding when a genetic algorithm for many-to-many goods scheduling is used for solving.

Description

Many-to-many goods scheduling optimization distribution method based on available resource coding
Technical Field
The invention belongs to the technical field of cargo scheduling, and particularly relates to a many-to-many cargo scheduling optimization distribution method.
Background
With the development of modern logistics technology, in goods transportation scheduling application, a problem of many-to-many goods scheduling application exists, and the so-called many-to-many goods scheduling means that a plurality of warehouses and a plurality of markets exist, a single warehouse supplies goods to the plurality of markets at the same time, and the goods in the single market can be delivered by the plurality of warehouses. As a large number of goods are constrained by certain passing probability due to factors such as height and width limitation, bearing and the like from a warehouse to a market road, the maximum success probability is taken as the optimization direction of a scheduling scheme in many-to-many goods scheduling application problems, and the objective function corresponding to the problems is in a nonlinear form. The genetic algorithm is widely applied as a calculator for solving the large-scale nonlinear combinatorial optimization problem, and the variable is not directly subjected to iterative operation, but feasible solutions are mapped into chromosome gene strings by using variable codes, and the globally optimal solution is searched by continuously performing genetic operation on the chromosome gene strings. However, when the genetic algorithm is applied to the problem of many-to-many goods scheduling, because a single warehouse goods is supplied to a plurality of markets at the same time, the single market goods can be delivered by a plurality of warehouses, and when the genetic operation is performed by the conventional numerical coding and re-cascading gene coding strategy, a plurality of constraint conditions are difficult to be considered at the same time, so that dead code phenomena such as goods type mismatching, unlimited available quantity, unsatisfied required quantity and the like easily occur. Therefore, how to utilize the genetic algorithm to solve the many-to-many goods scheduling optimization is a technical problem which needs to be solved urgently.
Disclosure of Invention
The invention aims to overcome the defects of the conventional technology, provides a many-to-many goods scheduling optimization allocation method based on available resource coding aiming at the complex constraint and objective function of many-to-many goods scheduling, solves the genetic dead code difficulty caused by conflict of a plurality of constraint conditions, and realizes the optimal solution of many-to-many goods scheduling by utilizing a genetic algorithm.
The invention has the following inventive concept: in the invention, the goods are mixedThe degree problem is abstractly described as follows: dispatching N-type cargos { N1, N2 and … Nn } from K warehouses { K1, K2 and … Kk } to M markets { M1, M2 and … Mm } for sale, and setting distribution variables
Figure 944760DEST_PATH_IMAGE001
Representing the number of allocations of the nth category of goods scheduled from the kth warehouse to the mth market, the known conditions are as follows:
a) demand situation of each kind of goods in each market
Figure 861900DEST_PATH_IMAGE002
Representing the demanded quantity of the nth type goods in the mth market;
b) availability of each type of goods per warehouse
Figure 172796DEST_PATH_IMAGE003
Representing the available quantity of the nth type goods in the kth warehouse;
c) probability of passable road for transporting certain goods from certain warehouse to certain market
Figure 738644DEST_PATH_IMAGE004
The probability of the trafficability of the road from the kth warehouse to the nth market is represented, and the probability is 0, so that the road is not trafficable;
the many-to-many allocation means: allowing a single warehouse to simultaneously supply multiple markets, but not exceeding the number of warehouses available, have
Figure 959541DEST_PATH_IMAGE005
(ii) a Allowing a single market to be delivered by multiple warehouses, but not less than the market demand, have
Figure 578741DEST_PATH_IMAGE006
。
The constraint to be considered in the many-to-many allocation comprises: the distribution scheme not only meets the requirements of the available quantity of the warehouse and the quantity of the market demand, but also can not generate matching items of the warehouse, the market and the goods type corresponding to the 0 element in the road passable probability matrix; i.e. the constraints include:
the optimization goal of the many-to-many distribution is the maximum success probability, namely the objective function is the success probability and the maximum value of completing various goods scheduling tasks, and is expressed as
Figure 517878DEST_PATH_IMAGE007
Solving by using genetic algorithm aiming at distribution variable due to non-linear expression in objective function
Figure 123303DEST_PATH_IMAGE008
Numerical encoding was performed in a cascade fashion, and individual genes were designed as follows:
Figure 588919DEST_PATH_IMAGE009
but in the constrained unbalanced many-to-many cargo scheduling problem described above,
Figure 487343DEST_PATH_IMAGE008
value range of (A) with subscriptknmThe number of the available warehouses, the number of the market demands, the road trafficability probability and other constraints exist, and an unfeasible solution can be generated by the unconstrained genetic operation. For example, in a crossover operator, if variable values are assigned at two points across any exchange of cargo types, the results of gene decoding may exceed the available number limit of the corresponding warehouse, and dead codes may be formed. The reason is that the conventional cascade numerical coding method describes the distribution relation by using variable subscripts, adopts numerical coding on variable values, and then serially splices a plurality of variables to form individual chromosome gene strings, is mostly applied to the problem of single-pair and single-assignment under no constraint, and only one warehouse of certain goods in one distribution is limited to supply goods to one market, so that the constraint condition of the expression form is avoided, and only the constraint of the variable value domain is met during population initialization.
Therefore, aiming at the complex constraint and objective function of many-to-many cargo scheduling, the basis is that when the genetic algorithm is used for solvingThe encoding method must simultaneously consider two dimensions of distribution relation and distribution quantity among warehouses, markets and goods. Therefore, the coding method based on the available resources is innovatively provided, and the distribution variables are reversed
Figure 913776DEST_PATH_IMAGE008
The upper limit of the original variable value domain is taken as the length of the chromosome gene string, the original variable value is represented by gene locus, the symbol coding is carried out on the variable subscript index, and the value symbol of the gene locus is used for representing the original variable subscript. The encoding method is utilized to obtain individual chromosome gene strings, and genetic operation is carried out to obtain an optimal distribution scheme.
The technical scheme of the invention is as follows: a many-to-many goods scheduling optimization distribution method based on available resource coding is characterized in that: the method comprises the steps of initializing a population according to the gene coding method, determining an individual fitness function, and performing selection, crossing and variation operations of a genetic algorithm to obtain an optimal solution of many-to-many goods scheduling;
the method specifically comprises the following steps: the method for carrying out gene coding based on the available resources of the warehouse goods comprises the following steps:
Step11 determining chromosome gene string form of segment after point
Sequentially listing all available standard unit resources corresponding to all goods types in the warehouse by taking the goods types in the warehouse as gene fragments and the available quantity of the goods as gene loci and taking the minimum distribution unit 1 as a standard, and determining the chromosome gene string form of the first section and the later section;
it can be seen that the length of the individual chromosome gene string is equal to the total amount of the goods available, and the step is to change a plurality of variable values
Figure 854050DEST_PATH_IMAGE010
Uniformly converting the non-zero value of the gene locus into an index of a fixed-length chromosome gene string;
step12 utilizes the probability matrix of the road passable symbol to generate the encoding matrix of the road passable symbol
Generating a road passable probability matrix according to the road passable probability of each warehouse for transporting each type of goods to each market;
(it is stated that each element value of the road passable probability matrix represents the passable probability of the corresponding goods in the corresponding warehouse reaching the corresponding market, and the corresponding element value of the two cases that no goods in a certain warehouse meet the requirement of the market or no goods in a certain warehouse meet the requirement of the market)
Coding a non-zero value part in the road passable probability matrix by using a non-0 symbol, and converting the road passable probability matrix into a road passable symbol coding matrix; the coding symbols establish an index relationship with the warehouse, the goods type and the market;
It can be seen from this that: the code symbol shows that a certain kind of goods in a certain kind of warehouse is allocated to a certain market, and the 0 code shows that the certain kind of goods in the certain kind of warehouse does not participate in allocation to the certain market; the road passable coding matrix realizes distribution variable
Figure 49540DEST_PATH_IMAGE008
Index of variable subscriptknmConverting the data into a symbol coding set of the gene locus value;
step13, obtaining selectable coding subsets of each gene segment by using the road trafficable probability coding matrix
Extracting a coding symbol set corresponding to each gene segment by using the road passable symbol coding matrix, adding a symbol 0 to represent the non-distribution condition, and obtaining a selectable coding subset of each gene segment;
description of the drawings: since each locus of the same gene fragment represents the same type of goods in the same warehouse, the selectable coding subsets of the coding symbols are the same.
Step14 initialization population
Randomly generating coding symbols of each gene locus in the selectable coding subset corresponding to each gene fragment determined in the Step13 according to the form of the chromosome gene string determined in the Step11 to obtain the chromosome gene string of the individual; repeating the steps for multiple times to obtain an initialized population;
the individual fitness function comprises a success probability and an objective function;
The cross operation is carried out among the same type of warehouse goods; the mutation operation is performed in an alternative coding subset of the gene segments.
Further, the non-zero value parts in Step12 are encoded sequentially by positive integers.
Further, the individual fitness function includes: using the target function of the success probability sum as an individual fitness function main body; and supplementing the quantity constraint condition meeting the market demand into the individual fitness function in the form of a penalty function.
Further, the crossing operation adopts multiple random two-point sequence crossing, and the specific steps are as follows: firstly, randomly determining the goods participating in the crossing as the nth goods, then extracting gene segments corresponding to the nth goods from two parent individuals A1 and B1, and splicing temporary gene strings A2 and B2 of the nth goods, thereby ensuring that the crossing is carried out in the nth goods; then two cross points p1 and p2 are randomly determined in the temporary gene strings A2 and B2, and the gene segments between the cross points p1 and p2 are exchanged point by point sequentially to obtain new temporary gene strings A3 and B3; and finally, returning the temporary genomes A3 and B3 to replace the gene positions of the original two parent individuals A1 and B1 to finally obtain two new child individuals A4 and B4.
Further, the variation operation adopts single-point uniform variation with a limited range, and comprises the following specific steps: randomly determining variation points in individual chromosome gene strings; the size of the variable point corresponding to the optional coding subset is
Figure 10542DEST_PATH_IMAGE011
Divide the (0,1) interval equally according to the subset size
Figure 422807DEST_PATH_IMAGE011
Equally dividing intervals, wherein each 1 interval corresponds to symbol codes of 1 optional code subset; taking the random numbers rand uniformly distributed in the (0,1) interval, and judging which interval the random numbers rand falls into, namely corresponding to the corresponding intervalThe symbol code of (2) mutates the point of variation.
Further, the selecting operation adopts a fitness proportion method to set the population scale as n, wherein the fitness value of the individual i is
Figure 229089DEST_PATH_IMAGE012
Then the probability that the individual i is selected is:
Figure 544664DEST_PATH_IMAGE013
in the formula (I), the compound is shown in the specification,
Figure 286355DEST_PATH_IMAGE014
the proportion of the fitness of the individual i in the total population fitness is reflected, and obviously, the higher the individual fitness is, the higher the probability of selection is.
Further, the individual fitness function is:
Figure 812014DEST_PATH_IMAGE015
in the formula:
Figure 297353DEST_PATH_IMAGE008
the distribution quantity from the nth goods to the mth market is scheduled from the kth warehouse, and the value of the distribution quantity is obtained by establishing an index relation between each genetic locus symbol code of an individual gene and the warehouse, the goods type and the market; k is the total number of warehouses, N is the total number of the dispatched goods categories, and M is the total number of markets for goods dispatching;
Figure 965970DEST_PATH_IMAGE016
For the penalty factor, if the distribution scheme meets the market demand quantity constraint, the individual fitness function will degenerate into an objective function with the maximum overall success probability, and if the distribution scheme does not meet the overall success probability, the corresponding fitness value is smaller than that of the objective function.
Further, an annealing operation is incorporated in the genetic algorithm operation, which specifically includes: after the individuals are subjected to selection, crossing and mutation operations to generate offspring individuals from parent individuals, the individuals are subjected to simulated annealing operation, and whether the parents or the offspring enter a next generation group is judged in a probability mode.
The invention has the beneficial effects that:
the method is characterized in that gene coding is carried out based on available resources, the form of segmenting and dividing chromosome gene strings corresponds to the available condition of each type of goods in each warehouse, the form of the chromosome gene strings is not changed by genetic operation, and the available quantity of the goods in the warehouse does not exceed the constraint and is always established;
on the basis of carrying out gene coding based on available resources, the invention provides correspondingly improved multiple random two-point sequence crossing operation and limited range single-point uniform variation operation which are both limited to be carried out in an optional coding subset, and the passable restriction of roads is ensured to be established; the designed fitness function comprises a penalty function representing the demand of the market goods, and when iterative operation is optimized, the demand quantity of the market goods can meet the constraint.
Drawings
FIG. 1 is a schematic flow chart of the present embodiment;
FIG. 2 is a schematic diagram of a chromosomal gene string of this example.
Detailed Description
Example 1
The technical scheme of the invention is further specifically described in the following description and specific examples in combination with the accompanying drawings.
As shown in fig. 1, a many-to-many goods scheduling optimization allocation method based on available resource coding is characterized by performing the following steps:
step1, carrying out gene coding based on available resources of warehouse goods, and initializing a population;
step2, determining an individual fitness function;
step3 individual carries on the selection, crossover, mutation operation, produces the filial generation individual by the father individual;
step4 individual carries out simulated annealing operation, and judges whether the parent generation or the offspring enters the next generation group;
step5 judges the termination condition, if not, returns to Step 3; and if the termination condition is met, outputting the current optimal individual as the optimal solution.
In order to clearly illustrate the method steps presented in the present invention, a specific example is provided to illustrate the nature of the method execution
The specific example assumes that 3 types of goods { apple, banana and orange } are scheduled for sale from 2 warehouses { warehouse 1, warehouse 2 } to 3 markets { east city, west city and south city }, the storage condition of the warehouse goods, the road passable probability and the market goods demand are known, and the optimal allocation scheme corresponding to the success probability and the maximum value is obtained.
The invention provides a many-to-many goods scheduling optimization allocation method based on available resource coding, which comprises the following steps:
step1 genetic codes are performed based on the available resources of the warehouse goods to initialize the population. The method comprises the following specific steps:
step11 determining chromosome gene string form of segment after point
The storage of the goods in the known warehouse is shown in the following table.
Figure 534354DEST_PATH_IMAGE017
The combination of warehouse and goods type has 5 cases of No. 1 storehouse apple, No. 1 storehouse banana, No. 2 storehouse apple, No. 2 storehouse banana, No. 2 storehouse orange, and the total number of gene segments is 5. According to the known available quantity of each type of goods in each warehouse, the number of the gene loci of each section is 1, 2, 3, 2 and 2 respectively, and the chromosome gene strings are obtained by sequentially expanding the sections and the points according to the nested relation as shown in figure 2.
Figure 891517DEST_PATH_IMAGE018
Step12, obtaining a road passable code matrix by using the road passable probability matrix
The probability matrix of the known road to pass (3 elements corresponding to apple, banana and orange respectively) is shown in the following table.
Figure 446126DEST_PATH_IMAGE019
Encoding a road trafficable probability matrix, wherein: the non-zero elements are encoded sequentially by positive integers, and the zero value parts are uniformly encoded with 0. The road trafficable probability code matrix is shown in the following table.
Figure 798610DEST_PATH_IMAGE020
The obtained symbol coding set {0,1,2,3,4,5,6,7,8} is a symbol set of possible values of all gene loci.
The probability matrix and the coding matrix of the road can pass through are equal to the distribution variable
Figure 272317DEST_PATH_IMAGE008
An index relationship is established, and the corresponding relationship is represented by the following table:
Figure 382356DEST_PATH_IMAGE021
step13, determining the selectable coding subset corresponding to each gene segment from the road passable coding matrix
And determining an optional subset of each gene segment by the road passable coding matrix aiming at the warehouse goods type represented by the gene segment, and combing optional markets or not participating in distribution conditions according to the warehouse goods type corresponding to the symbolic coding. For example, gene fragments
Figure 334131DEST_PATH_IMAGE022
Representing the apple bin 1, if there are four possible cases of no allocation, or allocation to east city, or allocation to west city, or allocation to south city, the corresponding selectable code subset is {0,1,3,4}, and repeating the process to obtain the selectable code subsets corresponding to all the gene segments in the example as shown in the following table.
Figure 977340DEST_PATH_IMAGE023
Step14 initializing population
And randomly determining a value in the range of the selectable coding subsets aiming at each gene locus, and obtaining an individual chromosome gene string. For example, have
Figure 497314DEST_PATH_IMAGE024
The resource allocation mode represented by each gene locus can be clearly known through the passable probability coding matrix, wherein
Figure 219282DEST_PATH_IMAGE025
Number 4 indicates that 1 unit of apple in bin No. 1 is allocated to south city;
Figure 115694DEST_PATH_IMAGE026
a number 2 indicates that bin No. 1, 1 unit of bananas, was allocated to east city;
Figure 583716DEST_PATH_IMAGE027
a 7 indicates that 1 unit of apple in bin No. 2 was allocated to western;
Figure 664804DEST_PATH_IMAGE028
a 6 denotes a number 2 bin with 1 unit banana allocated to east city;
Figure 77331DEST_PATH_IMAGE029
and
Figure 10390DEST_PATH_IMAGE030
each 8 indicates that 2 units of oranges in bin No. 2 were allocated to south city; because each gene locus only represents 1 unit resource, the distribution number of 2 needs to find two points in the corresponding gene segment to take the same symbol; the others are 0, indicating that the corresponding warehouse good is not involved in the resource allocation.
In order to keep the diversity of individuals in the population, the hamming distance is used as the criterion for initializing the generation of the population, the size of the population is G,estimating
Figure 864076DEST_PATH_IMAGE031
And C is the gene length given above, and 32 individual genes were generated when the species group was currently initialized.
Step2, integrating the objective function and the demand constraint of market goods, and calculating the individual fitness in the current population;
the individual fitness function comprises an objective function and a market cargo demand constraint, and is as follows:
Figure 584907DEST_PATH_IMAGE032
in the formula
Figure 156834DEST_PATH_IMAGE033
For penalty factors, if the allocation scheme meets the market demand quantity constraint, the individual fitness function will degrade to an objective function of the success probability sum.
The known market demand for goods is shown in the table below.
Figure 395049DEST_PATH_IMAGE034
Aiming at a chromosome gene string of a certain individual in an initialization population
Figure 227875DEST_PATH_IMAGE035
The corresponding allocation schemes are shown in the following table.
Figure 759089DEST_PATH_IMAGE036
Decoding to obtain
Figure 677366DEST_PATH_IMAGE037
=1、
Figure 719271DEST_PATH_IMAGE038
=1、
Figure 16392DEST_PATH_IMAGE039
=1、
Figure 610184DEST_PATH_IMAGE040
=1、
Figure 625545DEST_PATH_IMAGE041
And =2, taking values of 0.9, 0.8, 0.85, 0.9 and 0.95 by combining the road trafficability probability matrix, taking 0.1 as a penalty factor, and substituting 1 unit of apple in south city with a penalty function value of 1 into the fitness function to calculate F (x) =0.9 + 0.85+0.8 +0.9 +0.95 + 2-0.1 = 3.285.
Step3, selecting, crossing and mutating individuals, and generating child individuals from parent individuals, wherein the method specifically comprises the following steps:
the selection operation adopts a fitness proportion method (also called a roulette method), the group size is set as n, and the fitness value of an individual i is
Figure 969676DEST_PATH_IMAGE012
Then the probability that the individual i is selected is
Figure 777095DEST_PATH_IMAGE042
. In the formula (I), the compound is shown in the specification,
Figure 745051DEST_PATH_IMAGE014
the proportion of the fitness of the individual i in the total population fitness is reflected, and obviously, the higher the individual fitness is, the higher the probability of selection is.
The crossing operation adopts multiple random two-point sequence crossing, assuming that the two parent chromosome gene strings selected according to the fitness value are respectively A1= {1,2,0,0,7,0,6,0,8,8}, B1= {0,2,0,0,5,0,6,0,8,0}, firstly determining the type of the goods needing crossing to be apple by using random numbers, and then extracting the segments belonging to the apple in the two gene strings
Figure 247708DEST_PATH_IMAGE022
（
Figure 631416DEST_PATH_IMAGE043
) And
Figure 27762DEST_PATH_IMAGE044
(comprises
Figure 838723DEST_PATH_IMAGE045
) Spelling a temporary gene string to obtain a2= {1,0,7,0}, B2= {0,0,5,0 }; and then randomly determining that the two crossed points are 2 and 4, namely exchanging the symbol values between the 2 nd point and the 4 th point (including 2 and 4) of the temporary gene string, changing to A3= {1,0,5,0}, B3= {0,0,7,0}, finally returning to replace the original gene fragment position according to the extraction relation, and finally obtaining two new child chromosome gene strings A4= {1,2,0,0, 0,5,0,6,0,8,8}, B4= {0,2,0,0,7,0,6,0,8,0 }.
The mutation operation adopts a limited range of single-point uniform mutation, and if a parent chromosomal gene string P1= {1,2,0,0,7,0,6,0,8,8} is selected according to the fitness value, the mutation gene locus is randomly determined at first
Figure 953310DEST_PATH_IMAGE025
The optional coding subset is {0,1,3,4}, the equal division (0,1) interval 4 is equal division corresponding to the optional coding subset 0,1,3,4, if the random number is 0.2 and falls into the interval (0,0.25), the variation is symbol 0, and a new offspring chromosomal gene string P2= {0,2,0,0,7,0,6,0,8,8} is obtained.
Step4, performing simulated annealing operation to judge whether the parent generation or the offspring enters the next generation group, mainly using Metropolis receiving criterion, and setting the current problem
Figure 373664DEST_PATH_IMAGE046
Generating from the current state i an individual fitness value of
Figure 765463DEST_PATH_IMAGE012
After the operations of selection, crossing, mutation and the like of the genetic algorithm, the new individual fitness under the new state j is generated as
Figure 871959DEST_PATH_IMAGE012
Setting the fitness variance as
Figure 614787DEST_PATH_IMAGE047
The transition probability p corresponding to the Metropolis reception criterion obeys the following distribution:
Figure 74718DEST_PATH_IMAGE048
the method comprises the following specific steps: if it is not
Figure 711236DEST_PATH_IMAGE049
Then new individual fitness value at state j
Figure 96956DEST_PATH_IMAGE050
The corresponding individual is retained in the next generation and the fitness is adjusted
Figure 61501DEST_PATH_IMAGE051
The corresponding individual is deleted from the population; if it is not
Figure 449757DEST_PATH_IMAGE052
Calculating the corresponding transition probability p according to the formula; when in use
Figure 144043DEST_PATH_IMAGE053
New individual fitness value at state j at random number
Figure 467708DEST_PATH_IMAGE054
The corresponding individual is retained in the next generation and the fitness is adjusted
Figure 716287DEST_PATH_IMAGE012
The corresponding individual is deleted from the population; when in use
Figure 111496DEST_PATH_IMAGE055
New individual fitness value at state i at random number
Figure 660289DEST_PATH_IMAGE051
The corresponding individual is retained in the next generation and the fitness is adjusted
Figure 918970DEST_PATH_IMAGE050
The corresponding individual is deleted from the population.
Step5, judging a termination condition, if the termination condition is not met, executing a temperature-reducing operation, and returning to Step 2; and if the termination condition is met, outputting the most optimal solution of the current optimal individual. With specific examples, the following parameters were employed: the population average fitness change threshold is 0.1, the maximum genetic algebra is 200, the initial temperature is 1000, the cut-off low temperature is 1, the cooling rate is 0.95, and the example approximate optimal solution is
Figure 717162DEST_PATH_IMAGE056
The success probability and the maximum value are 6.25, namely 1 unit of apple in bin 1 is allocated to south city, 1 unit of apple in bin 2 is allocated to east city, 1 unit of apple in bin 2 is allocated to west city, 2 unit of banana in bin 2 is allocated to east city, and 2 unit of orange in bin 2 is allocated to south city.

Claims (6)

1. A many-to-many goods scheduling optimization distribution method based on available resource coding is characterized in that: the method comprises the steps of initializing a population according to the gene coding method, determining an individual fitness function, and iteratively performing selection, intersection and variation operations of a genetic algorithm to obtain an optimal solution of many-to-many goods scheduling; the method specifically comprises the following steps:
the gene coding method based on the available resources of the warehouse goods comprises the following steps:
step11 determining chromosome gene string form of segment and point
Sequentially listing all available standard unit resources corresponding to all goods types in the warehouse by taking the goods types in the warehouse as gene fragments and the available quantity of the goods as gene loci and taking the minimum distribution unit 1 as a standard, and determining the chromosome gene string form of the first section and the later section;
step12 utilizes the probability matrix of the road passable symbol to generate the encoding matrix of the road passable symbol
Generating a road passable probability matrix according to the road passable probability of each warehouse for transporting each type of goods to each market; each element value of the road passable probability matrix is the passable probability of various goods in each warehouse reaching each market, and 0 is taken for the corresponding element value of two conditions that no goods in a certain warehouse meet the requirement of the market or no goods in the certain warehouse meet the requirement of the market in the certain market;
coding a non-zero value part in the road passable probability matrix by using a non-0 symbol, and converting the road passable probability matrix into a road passable symbol coding matrix; defining the non-0 symbol as a coding symbol, wherein the coding symbol establishes an index relationship with a warehouse, a cargo type and a market;
step13 uses the road passable symbol coding matrix to obtain the selectable coding subset of each gene segment
Extracting a coding symbol set corresponding to each gene segment by using the road passable symbol coding matrix, adding a symbol 0 to represent the non-distribution condition, and obtaining a selectable coding subset of each gene segment;
step14 initialization population
Randomly generating coding symbols of each gene locus in the selectable coding subset corresponding to each gene fragment determined in the Step13 according to the form of the chromosome gene string determined in the Step11 to obtain the chromosome gene string of the individual; repeating the steps for multiple times to obtain an initialized population;
The individual fitness function comprises a success probability and an objective function;
the cross operation is carried out among the same type of warehouse goods; the mutation operation is performed in an alternative coding subset of the gene fragment.
2. The method as claimed in claim 1, wherein the method for optimal allocation of the scheduling of many-to-many goods based on the available resource codes comprises: the non-zero value part is sign-coded by a positive integer in Step 12.
3. The method as claimed in claim 1, wherein the individual fitness function comprises: using the target function of the success probability sum as an individual fitness function main body; and supplementing the quantity constraint condition meeting the market demand into the individual fitness function in the form of a penalty function.
4. A many-to-many optimal distribution method for cargo scheduling based on available resource coding as claimed in claim 1, wherein said interleaving operation employs multiple random two-point sequential interleaving, and comprises the following specific steps: firstly, randomly determining the goods participating in the crossing as nth goods, then extracting gene segments corresponding to the nth goods from two parent individuals A1 and B1, and splicing temporary gene strings A2 and B2 of the nth goods, thereby ensuring that the crossing is carried out in the nth goods; then, two cross points p1 and p2 are randomly determined in the temporary gene strings A2 and B2, and gene segments between the cross points p1 and p2 are exchanged point by point sequentially to obtain new temporary gene strings A3 and B3; and finally, returning the temporary gene strings A3 and B3 to replace the gene positions of the original two parent individuals A1 and B1, and finally obtaining two new child individuals A4 and B4.
5. The method as claimed in claim 1, wherein the mutation operation employs a limited range of single-point uniform mutation, and comprises the following steps: randomly determining variation points in the chromosome gene string; the size of the variable point corresponding to the optional coding subset is
Figure 745652DEST_PATH_IMAGE001
Divide the (0,1) interval equally according to the subset size
Figure 393671DEST_PATH_IMAGE002
Equally dividing intervals, wherein each 1 interval corresponds to symbol codes of 1 optional code subset; taking the random number rand uniformly distributed in the (0,1) interval, judging which interval the random number rand falls into, namely using the symbol code corresponding to the corresponding interval to carry out the variation pointAnd (5) carrying out mutation.
6. A many-to-many cargo scheduling optimized allocation method based on available resource coding as claimed in claim 1, characterized in that the genetic algorithm operation incorporates an annealing operation, specifically: after the individuals are subjected to selection, crossing and mutation operations, filial individuals are generated by parent individuals, the individuals are subjected to simulated annealing operation, and whether the parents or the filial individuals enter a next generation group is judged in a probability mode.
CN202210352848.5A 2022-04-06 2022-04-06 Many-to-many goods scheduling optimization distribution method based on available resource coding Active CN114493047B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210352848.5A CN114493047B (en) 2022-04-06 2022-04-06 Many-to-many goods scheduling optimization distribution method based on available resource coding

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210352848.5A CN114493047B (en) 2022-04-06 2022-04-06 Many-to-many goods scheduling optimization distribution method based on available resource coding

Publications (2)

Publication Number Publication Date
CN114493047A CN114493047A (en) 2022-05-13
CN114493047B true CN114493047B (en) 2022-06-14

Family

ID=81487744

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210352848.5A Active CN114493047B (en) 2022-04-06 2022-04-06 Many-to-many goods scheduling optimization distribution method based on available resource coding

Country Status (1)

Country Link
CN (1) CN114493047B (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107578119A (en) * 2017-08-09 2018-01-12 泉州装备制造研究所 A kind of resource allocation global optimization method of intelligent dispatching system
CN111178582A (en) * 2019-11-29 2020-05-19 浙江工业大学 Logistics distribution optimization method based on improved genetic algorithm
CN112884257A (en) * 2021-04-29 2021-06-01 牧星机器人(江苏)有限公司 Goods taking path optimization method, device and system based on genetic algorithm

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
IT201700056428A1 (en) * 2017-05-24 2018-11-24 Telespazio Spa INNOVATIVE SATELLITE SCHEDULING METHOD BASED ON GENETIC ALGORITHMS AND SIMULATED ANNEALING AND RELATIVE MISSION PLANNER

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107578119A (en) * 2017-08-09 2018-01-12 泉州装备制造研究所 A kind of resource allocation global optimization method of intelligent dispatching system
CN111178582A (en) * 2019-11-29 2020-05-19 浙江工业大学 Logistics distribution optimization method based on improved genetic algorithm
CN112884257A (en) * 2021-04-29 2021-06-01 牧星机器人(江苏)有限公司 Goods taking path optimization method, device and system based on genetic algorithm

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
多受灾点应急救援车辆调度的优化遗传算法;喻德旷等;《计算机系统应用》;20161115(第11期);全文 *

Also Published As

Publication number Publication date
CN114493047A (en) 2022-05-13

Similar Documents

Publication Publication Date Title
CN108320057B (en) Flexible job shop scheduling method based on limited stable pairing strategy
CN109919541B (en) Modeling solving method for multi-stage positioning inventory path problem
CN111507641B (en) Batch processing equipment scheduling method and device
Luo et al. Improved shuffled frog leaping algorithm and its multi-phase model for multi-depot vehicle routing problem
CN105096006A (en) Method for optimizing a routing of an intelligent ammeter distributing vehicle
CN109800910B (en) Vehicle route optimization method based on tabu search hyperheuristic algorithm
CN107578119A (en) A kind of resource allocation global optimization method of intelligent dispatching system
CN106682769A (en) Method and apparatus for realizing product vehicle route
CN109784530B (en) Power supply partition optimization setting method based on medium-speed magnetic suspension line passing capacity
CN109886490A (en) A kind of intermodal matching optimization method of combined vehicle
CN110807559A (en) Order batching and picking path combined optimization method
CN111368999B (en) Logistics scheduling system
CN102663499B (en) Network community division method based on simulated annealing genetic algorithm
CN113033895A (en) Multi-source multi-point path planning method, equipment and storage medium
CN111027911B (en) Automatic logistics storage transportation scheduling system
CN113570025B (en) E-commerce storage center goods space distribution method based on discrete particle swarm optimization
CN114897217A (en) Fresh food electricity merchant front-end bin site selection-path planning method
CN103927628A (en) Order management system and order management method oriented to customer commitments
CN102298737A (en) Customer commitment-oriented order management system and method thereof
CN111199375A (en) Intelligent logistics transportation system
CN114493047B (en) Many-to-many goods scheduling optimization distribution method based on available resource coding
CN117252395B (en) Double-chromosome genetic algorithm-based multi-logistics vehicle scheduling method with service constraint
CN115345549A (en) Vehicle path adjusting method and system combined with loading scheme
CN116341860A (en) Vehicle dispatching optimization method based on improved NSGA-II algorithm
WO2024022054A1 (en) Flexible job-shop production scheduling method and apparatus

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant