A logistics scheduling management system based on big data

Abstract

The present invention relates to the technical field of scheduling management, and discloses a logistics scheduling management system based on big data, including a data acquisition module, a data classification module, a scheme initial screening module, a scheme optimization module and an intelligent management and control module; by providing the data classification module, the scheme initial screening module and the scheme optimization module, it is beneficial to obtain marked data categories by classifying the data in the real-time logistics data set, and can match the scheduling content for the marked data category, while not matching the unlabeled data category, thereby reducing the amount of data calculation, improving efficiency, and improving accuracy, ensuring that all scheduling contents comply with the latest laws and regulations and policies, and forming an alternative scheduling scheme set based on the matched scheduling content, which lays a foundation for subsequent acquisition of the best logistics scheduling scheme, and the scheme optimization module combines a genetic algorithm with a bat algorithm, so as to be more conducive to obtaining the best logistics scheduling scheme from complex and changeable scheduling schemes.

Description

Technical Field

[0001] The present invention relates to the technical field of scheduling management, and more specifically to a logistics scheduling management system based on big data. Background Art

[0002] Logistics scheduling management aims to provide enterprises with efficient, accurate and intelligent logistics scheduling solutions through big data analysis, artificial intelligence algorithms and Internet of Things technology. Logistics scheduling involves multiple aspects and is also affected by many aspects. The impact of policies and regulations on logistics scheduling is highly valued by logistics companies; for example, restrictions on vehicle weight and size will affect the loading and transportation methods of goods, and scheduling strategies need to be adjusted; the existing technology has a variety of analyses on logistics scheduling management, which optimizes logistics scheduling management from various angles and improves the intelligence level of logistics scheduling management.

[0003] However, the existing technology lacks the analysis of policies and regulations to adjust the corresponding logistics scheduling plans, and is unable to capture and analyze policies and regulations in a timely manner, which will cause logistics companies to reduce logistics transportation efficiency and logistics company benefits, and have impacts on logistics companies in many aspects including cost, efficiency, and safety. Therefore, it is very important to analyze the changes in policies and regulations in a timely manner and adopt corresponding logistics scheduling plans to improve the benefits of logistics companies. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a logistics scheduling management system based on big data to solve the problems existing in the above-mentioned background technology.

[0005] The present invention provides the following technical solutions: a logistics dispatching management system based on big data, comprising a data collection module, a data classification module, a solution initial screening module, a solution optimization module and an intelligent management and control module;

[0006] The data collection module is used to collect various historical logistics data sets in advance and collect real-time logistics data sets in real time;

[0007] The data classification module is used to classify various types of historical logistics data sets collected in advance and divide them into N data clusters; and classify the n text data in the real-time logistics data set collected in real time respectively to obtain the marked data category;

[0008] The scheme initial screening module is used to perform feasibility matching on the original logistics scheduling content and form a set of alternative scheduling schemes;

[0009] The scheme optimization module is used to obtain the best logistics scheduling scheme from the set of alternative scheduling schemes by combining the optimization algorithm;

[0010] The intelligent management and control module is used to execute the optimal logistics scheduling plan and to intelligently manage and control the logistics scheduling process.

[0011] Preferably, the data categories are types of logistics data involved in the logistics scheduling process, including vehicles, routes, goods, and personnel configuration;

[0012] The feasibility matching is specifically to match the scheduling content data of the original logistics scheduling solution set with the text content data in the real-time logistics data set, and obtain a set of alternative scheduling solutions that meet the conditions based on the matching results.

[0013] Preferably, the N data clusters are obtained in the following manner:

[0014] Setting digital labels for the data in the historical logistics data set and the real-time logistics data set;

[0015] Based on the text data features in the historical logistics data set, the content similarity matrix U is constructed. , where comm(x,y) represents the content similarity between the text data with the digital label x and the text data with the digital label y; the formula is: ;

[0016] Among them, k 1 is the weight parameter, zx is the feature vector of the text data with the digital label x, and zy is the feature vector of the text data with the digital label y; cov(zx,zy) is the covariance of zx and zy, σ(zx) is the standard deviation of zx, and σ(zy) is the standard deviation of zy;

[0017] Based on the content similarity matrix, the data in the historical logistics data set are clustered; each text data is regarded as a separate small cluster, and there are Q small clusters in total, where Q is the total number of text data in the historical logistics data set; the gravity value between each two small clusters is calculated based on the content similarity matrix; a merging threshold b is preset to control cluster merging, and the small clusters are iteratively merged. In each iteration, two small clusters q1 and q2 with the largest gravity value are found;

[0018] If the gravity value between q1 and q2 is greater than the merging threshold b, q1 and q2 are merged into a new cluster q, and the gravity value between the new cluster q and all other small clusters is recalculated; otherwise, no iterative merging is performed; the execution is repeated until the gravity value between no two small clusters is greater than the merging threshold b, at which time N new clusters are obtained, that is, N data clusters.

[0019] Preferably, the calculation formula of the gravity value is expressed as: ; Where Y(q1,q2) is the gravitational value between q1 and q2, G is the gravitational constant, c1 is the number of text data in small cluster q1, c2 is the number of text data in small cluster q2, Dis(q1,q2) is the distance between small clusters q1 and q2; that is, the average value of the distance between all text data pairs in the two small clusters; ω is a hyperparameter, σ(q1) is the standard deviation of the feature vector of the text data in small cluster q1, σ(q2) is the standard deviation of the feature vector of the text data in small cluster q2, μ(q1) is the mean of the feature vector of the text data in small cluster q1, and μ(q2) is the mean of the feature vector of the text data in small cluster q2.

[0020] Preferably, the method of obtaining the labeled data category is:

[0021] Calculate the gravity values ​​of n text data in the real-time logistics data set collected in real time and N data clusters respectively, arrange the gravity values ​​of the i-th text data and the N data clusters from large to small, select the data cluster corresponding to the maximum gravity value as the data category of the i-th text data, and mark the data category to obtain m marked data categories, each marked data category can correspond to multiple text data in the real-time logistics data set; i=1, 2, 3, ..., n; and classify the i-th text data into the data category; , where Y(i,j) is the attraction value between the i-th text data and the j-th data cluster, Cj is the number of text data in the j-th data cluster, Dis(i,j) is the distance between the i-th text data and the j-th data cluster; that is, the average value of the distance between the i-th text data and all the data in the j-th data cluster; σ(i) is the standard deviation of the eigenvector of the i-th text data, σ(j) is the standard deviation of the eigenvector of the text data in the j-th data cluster, μ(i) is the mean of the eigenvector of the i-th text data, μ(j) is the mean of the eigenvector of the text data in the j-th data cluster; j=1, 2, 3, …, N.

[0022] Preferably, the scheduling scheme set is composed of multiple scheduling schemes, each data category corresponds to a scheduling content set, the scheduling content set corresponding to the marked data category is all scheduling contents, and the scheduling content set corresponding to the unmarked data category is the scheduling content after the last feasibility matching of the category; all the contents in the scheduling content set are combined one by one to form all the scheduling schemes;

[0023] The original logistics scheduling content is the scheduling content corresponding to the marked data category and the scheduling content corresponding to the unmarked data category. The specific method of forming the alternative scheduling solution set is:

[0024] Obtain m labeled data categories, perform feasibility matching on the scheduling contents corresponding to the labeled data categories, and obtain a new scheduling content set after matching. The scheduling contents in the new scheduling content set are the scheduling contents after feasibility matching.

[0025] Feasibility matching includes:

[0026] Match the dispatch content corresponding to the marked data category with the text data classified in the corresponding data category; use keyword matching and semantic analysis or natural language processing related technologies to match; if a dispatch content does not match or conflicts with any text data classified in the corresponding data category, the matching result of the dispatch content is a match failure; if a dispatch content matches all text data classified in the corresponding data category and there is no conflict, the matching result of the dispatch content is a match success;

[0027] All successfully matched scheduling contents constitute a new scheduling content set; all new scheduling content sets are combined one by one with the contents of the scheduling content set corresponding to the non-labeled data category to form all alternative scheduling schemes, and the set composed of all alternative scheduling schemes is the alternative scheduling scheme set.

[0028] Preferably, the specific method of obtaining the optimal logistics scheduling solution is:

[0029] Step S01: Encode the alternative scheduling schemes in the alternative scheduling scheme set as a, randomly select R to construct an initial population, and initialize all parameters; the initial iteration number λ is 0;

[0030] Step S02: determine the fitness function; when the number of iterations is plural, execute step S03; when the number of iterations is singular, execute step S04;

[0031] Step S03: perform multiple iterations, and then directly execute step S05;

[0032] Step S04: perform a single number iteration, and then directly execute step S05;

[0033] Step S05: Determine whether the maximum number of iterations has been reached. If so, the iteration ends, and the alternative scheduling scheme corresponding to the maximum fitness value or the optimal position is obtained as the optimal logistics scheduling scheme; otherwise, let λ=λ+1, and loop from step S02 to step S05;

[0034] The method of judging whether the maximum number of iterations is reached is: when the fitness is greater than or equal to the fitness threshold Q or converges to a stable value, it is judged that the maximum number of iterations is reached.

[0035] Preferably, the initial population is expressed as: , where A represents the initial population, ar is the rth alternative scheduling scheme, r=1, 2, 3, …, R;

[0036] The fitness function is expressed as: , where f r is the fitness of the rth alternative scheduling solution, COS r is the cost of the rth alternative scheduling solution.

[0037] Preferably, step S03 specifically includes:

[0038] Step S031: consider a as a chromosome; a r is the rth chromosome, corresponding to the rth alternative scheduling scheme;

[0039] Step S032: performing natural selection on chromosomes in the population; the natural selection is performed by combining the elite method and the rotation method;

[0040] Step S033: performing crossover recombination on chromosomes in the population;

[0041] Step S034: mutate the chromosomes in the population to obtain a new population.

[0042] Preferably, the step S04 specifically includes:

[0043] Step S041: The population size is R, each alternative scheduling scheme can be regarded as a bat, and the position of the rth bat is p r , the speed is v r , the sound wave frequency is H r , the loudness of the sound wave is B r , the pulse frequency is I r ;

[0044] Step S042: Find the optimal position p* of the bats in the current population, and update the bat position and speed;

[0045] The bat position update formula is: , where p r λ is the position of the rth bat at the λth iteration, p r λ-1 is the position of the rth bat at the λ-1th iteration; v r λ is the speed of the rth bat at the λth iteration;

[0046] The bat speed update formula is: , where v r λ-1 is the speed of the rth bat at the λ-1th iteration, H ris the frequency of the sound wave emitted by the rth bat;

[0047] , where H min is the minimum frequency of the bat's sound waves, H max is the maximum frequency of the sound waves generated by bats, and the sound wave frequency range is [H min , H max ], β is a random vector between [0,1];

[0048] Step S043: Generate a random number rand1 between [0, 1]. If rand1>I r , then random flight is performed, and a new position p is generated near the original optimal position by random flight new , otherwise update the bat position according to the bat position update formula; the formula for random flight is:

[0049] , where p old is the original optimal position, rand2 is a random number between [0, 1], B λ is the average loudness of all bats in the λth iteration;

[0050] Step S044: Generate another random number rand3 on [0, 1]. If rand3 < B r , and the fitness corresponding to the new optimal position is greater than the fitness of the original optimal position, then the position is accepted and the new position is used as the current optimal position, and the sound wave loudness is reduced and the pulse frequency is increased;

[0051] The formula for adjusting the loudness and frequency of sound waves is:

[0052] ; ;in, , is the sound wave loudness attenuation coefficient, γ>0, is the pulse frequency enhancement coefficient, B r λ+1 is the sound wave loudness of the rth bat at the λ+1th iteration, B r λ is the sound wave loudness of the rth bat at the λth iteration, I r λ+1 is the pulse frequency of the rth bat at the λ+1th iteration, I r 0 is the initial pulse frequency of the rth bat.

[0053] Technical effects and advantages of the present invention:

[0054] The present invention is provided with a data classification module, a scheme initial screening module and a scheme recommendation module, which is conducive to obtaining marked data categories by classifying data in a real-time logistics data set, and can match the scheduling content for the marked data categories, while not matching the unlabeled data categories, thereby reducing the amount of data calculation, improving efficiency, and at the same time improving accuracy, ensuring that all scheduling contents comply with the latest laws and policies, and forming a set of alternative scheduling schemes based on the matched scheduling contents, laying the foundation for subsequent acquisition of the best logistics scheduling scheme. The scheme recommendation module combines the genetic algorithm with the bat algorithm. The genetic algorithm is good at global search, while the bat algorithm is good at local search. The combination of the two can enhance the local search capability lacking in the genetic algorithm and the global search capability lacking in the bat algorithm. The genetic algorithm and the bat algorithm can share the information of the optimal solution, so that both algorithms can benefit from each other's search results, improve the efficiency and accuracy of solving problems, and thus be more conducive to obtaining the best logistics scheduling scheme from complex and changeable scheduling schemes. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a structural diagram of the logistics scheduling management system based on big data of the present invention. DETAILED DESCRIPTION

[0056] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. In addition, the forms of the various structures recorded in the following embodiments are merely illustrative. The logistics scheduling management system based on big data involved in the present invention is not limited to the various structures recorded in the following embodiments. All other implementations obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0057] like Figure 1 As shown, the present invention provides a logistics scheduling management system based on big data, including a data collection module, a data classification module, a solution initial screening module, a solution optimization module and an intelligent management and control module;

[0058] The data collection module is used to pre-collect various historical logistics data sets and real-time logistics data sets; the collection method can be obtained through web crawler technology combined with machine learning models, the web crawler technology is used to collect all policy and regulatory text data, and the machine learning model is used to extract content with similar features from the collected policy and regulatory texts, and can also perform information extraction to automatically extract structured information from the text, such as events, relationships, etc.; the machine learning model can be trained through various historical logistics data collected in advance to extract text content related to logistics data in policy and regulatory text data, and can also be combined with setting keywords to improve the accuracy of logistics data extraction; the logistics data set is composed of multiple text data, each of which is a text content related to logistics transportation in a policy and regulatory text, including but not limited to the content of logistics transportation method regulations, logistics transportation tool regulations, and cargo loading weight and size regulations, etc.;

[0059] The data classification module is used to classify various types of historical logistics data sets collected in advance, and divide them into N data clusters, each data cluster represents a data category, and the sample points in each data cluster are text contents related to the data category; and the n text data in the real-time logistics data set collected in real time are classified respectively to obtain the marked data category; the data category is the type of logistics data involved in the logistics scheduling process, including vehicles, routes, goods and personnel configuration, etc.; illustratively, if the data category is vehicle, then the data content involved in the data category is all the text content of laws and regulations related to vehicle requirements in the logistics transportation process, such as requirements for the environmental protection of vehicles, requirements for vehicle speed, etc.; if the data category is route, then the data content involved in the data category is all the text content of laws and regulations related to the transportation route, such as restricting a certain section of road, temporarily controlling a certain section of road, etc.;

[0060] The scheme initial screening module is used to perform feasibility matching on the original logistics scheduling content and form a set of alternative scheduling schemes; the feasibility matching is specifically to match the scheduling content data of the original logistics scheduling scheme set with the text content data in the real-time logistics data set, and obtain a set of alternative scheduling schemes that meet the conditions based on the matching results;

[0061] The scheme optimization module is used to obtain the best logistics scheduling scheme from the set of alternative scheduling schemes by combining the optimization algorithm; the combined optimization algorithm is an optimization algorithm formed by combining the genetic algorithm and the bat algorithm. The genetic algorithm is good at global search, while the bat algorithm is good at local search. The combination of the two can enhance the local search ability lacking in the genetic algorithm and the global search ability lacking in the bat algorithm, and use the local search ability of the bat algorithm to accelerate convergence to the global optimal solution. The combined optimization algorithm shows better performance in processing complex optimization problems of logistics scheduling, and can obtain the best logistics scheduling scheme from multiple logistics scheduling schemes;

[0062] The intelligent management and control module is used to execute the optimal logistics scheduling plan and to intelligently manage and control the logistics scheduling process.

[0063] In this embodiment, it should be specifically explained that the method for obtaining the N data clusters is specifically as follows:

[0064] Setting digital labels for the data in the historical logistics data set and the real-time logistics data set;

[0065] Based on the text data features in the historical logistics data set, the content similarity matrix U is constructed. , where comm(x,y) represents the content similarity between the text data with the digital label x and the text data with the digital label y; the formula is: ;

[0066] Among them, k 1 is the weight parameter, zx is the feature vector of the text data with the digital label x, and zy is the feature vector of the text data with the digital label y; cov(zx,zy) is the covariance of zx and zy, σ(zx) is the standard deviation of zx, and σ(zy) is the standard deviation of zy;

[0067] Based on the content similarity matrix, the data in the historical logistics data set are clustered; each text data is regarded as a separate small cluster, and there are Q small clusters in total, where Q is the total number of text data in the historical logistics data set; the gravity value between each two small clusters is calculated based on the content similarity matrix; a merging threshold b is preset to control cluster merging, and the small clusters are iteratively merged. In each iteration, two small clusters q1 and q2 with the largest gravity value are found;

[0068] If the gravity value between q1 and q2 is greater than the merging threshold b, q1 and q2 are merged into a new cluster q, and the gravity value between the new cluster q and all other small clusters is recalculated; otherwise, no iterative merging is performed; the execution is repeated until the gravity value between no two small clusters is greater than the merging threshold b, at which time N new clusters are obtained, that is, N data clusters;

[0069] By adjusting the size of the merge threshold b, the strictness of the merge can be controlled, thereby affecting the final number of clusters. The size of the merge threshold b needs to be determined according to actual needs to determine a reasonable merge threshold so that the ideal number of clusters N is finally obtained.

[0070] In this embodiment, it should be specifically explained that the calculation formula of the gravity value is expressed as: ; Where Y(q1,q2) is the gravitational value between q1 and q2, G is the gravitational constant, c1 is the number of text data in small cluster q1, c2 is the number of text data in small cluster q2, Dis(q1,q2) is the distance between small clusters q1 and q2; that is, the average value of the distance between all text data pairs in the two small clusters; ω is a hyperparameter, σ(q1) is the standard deviation of the feature vector of the text data in small cluster q1, σ(q2) is the standard deviation of the feature vector of the text data in small cluster q2, μ(q1) is the mean of the feature vector of the text data in small cluster q1, and μ(q2) is the mean of the feature vector of the text data in small cluster q2.

[0071] In this embodiment, it should be specifically explained that the method for obtaining the marked data category is:

[0072] Calculate the gravity values ​​of n text data in the real-time logistics data set collected in real time and N data clusters respectively, arrange the gravity values ​​of the i-th text data and the N data clusters from large to small, select the data cluster corresponding to the maximum gravity value as the data category of the i-th text data, and mark the data category to obtain m marked data categories, each marked data category can correspond to multiple text data in the real-time logistics data set; i=1, 2, 3, ..., n; and classify the i-th text data into the data category; , where Y(i,j) is the attraction value between the i-th text data and the j-th data cluster, Cj is the number of text data in the j-th data cluster, Dis(i,j) is the distance between the i-th text data and the j-th data cluster; that is, the average value of the distance between the i-th text data and all the data in the j-th data cluster; σ(i) is the standard deviation of the eigenvector of the i-th text data, σ(j) is the standard deviation of the eigenvector of the text data in the j-th data cluster, μ(i) is the mean of the eigenvector of the i-th text data, μ(j) is the mean of the eigenvector of the text data in the j-th data cluster; j=1, 2, 3, …, N.

[0073] In this embodiment, it should be specifically explained that the scheduling plan set is composed of multiple scheduling plans, and the specific contents of the scheduling plan include vehicle tools, transportation routes, cargo loading weight and accompanying personnel, etc. Each data category corresponds to a scheduling content set, the scheduling content set corresponding to the marked data category is all scheduling contents, and the scheduling content set corresponding to the unmarked data category is the scheduling content after the last feasibility matching of the category; the contents in all scheduling content sets are combined one by one to form all scheduling plans; for example, if the scheduling content set corresponding to the vehicle tool is {small truck, medium truck}, the scheduling content set corresponding to the transportation route is {route a, route b}, the content set corresponding to the cargo loading weight is {50KG, 60KG}, and the scheduling content set corresponding to the accompanying personnel is {1 person, 2 people}, then the contents in the scheduling content set are combined one by one to form a scheduling. The plans include {small truck, route a, 50KG, 1 person}, {small truck, route b, 50KG, 1 person}, {small truck, route a, 60KG, 1 person}...; the scheduling content set can be obtained by collecting historical data corresponding to each data category, and combining the professional knowledge of technical personnel in this field and the actual possible logistics scheduling arrangements; the transportation route can be obtained according to the destination set in advance, so different logistics scheduling plans can be selected for different destinations; since the non-marked data category has not undergone content updates of laws and regulations, and only when the laws and regulations change, will the marked data category be formed and feasibility matching will be performed. Therefore, the scheduling content set of the non-marked data category selects the scheduling content after the last feasibility match, that is, the scheduling content corresponding to the latest policy, ensuring that all scheduling contents are within the scope permitted by laws and regulations.

[0074] The original logistics scheduling content is the scheduling content corresponding to the marked data category and the scheduling content corresponding to the unmarked data category. The specific method of forming the alternative scheduling solution set is:

[0075] Obtain m labeled data categories, perform feasibility matching on the scheduling contents corresponding to the labeled data categories, and obtain a new scheduling content set after matching. The scheduling contents in the new scheduling content set are the scheduling contents after feasibility matching.

[0076] Feasibility matching includes:

[0077] Match the dispatch content corresponding to the marked data category with the text data classified in the corresponding data category; use keyword matching and semantic analysis or natural language processing related technologies to match; if there is a dispatch content that does not match or conflicts with any text data classified in the corresponding data category, the matching result of the dispatch content is a match failure; this means that the dispatch content is different from the intention or instruction expressed by the text data, or even has an opposite intention or instruction, and therefore does not comply with the regulations and policies represented by the text data; if there is a dispatch content that matches all the text data classified in the corresponding data category and there is no conflict, the matching result of the dispatch content is a match success; this means that the dispatch content is similar to the intention or instruction expressed by all the text data in the data category and complies with the regulations and policies represented by the data category;

[0078] All successfully matched scheduling contents constitute a new scheduling content set; all new scheduling content sets are combined one by one with the contents of the scheduling content set corresponding to the non-labeled data category to form all alternative scheduling schemes, and the set composed of all alternative scheduling schemes is the alternative scheduling scheme set.

[0079] In this embodiment, it should be specifically explained that the specific method of obtaining the optimal logistics scheduling plan is:

[0080] Step S01: Encode the alternative scheduling schemes in the alternative scheduling scheme set as a, randomly select R to construct an initial population, and initialize all parameters; the initial iteration number λ is 0;

[0081] Step S02: determine the fitness function; when the number of iterations is plural, execute step S03; when the number of iterations is singular, execute step S04;

[0082] Step S03: perform multiple iterations, and then directly execute step S05;

[0083] Step S04: perform a single number iteration, and then directly execute step S05;

[0084] Step S05: Determine whether the maximum number of iterations has been reached. If so, the iteration ends, and the alternative scheduling scheme corresponding to the maximum fitness value or the optimal position is obtained as the optimal logistics scheduling scheme; otherwise, let λ=λ+1, and loop from step S02 to step S05; each completion of step S02 to step S05 is regarded as one iteration;

[0085] The method of determining whether the maximum number of iterations has been reached is: when the fitness is greater than or equal to the fitness threshold Q or converges to a certain stable value, it is determined that the maximum number of iterations has been reached;

[0086] The fitness threshold Q is preset by technical personnel in this field according to the accuracy of the algorithm. The population algebra is obtained by technical personnel in this field under multiple different alternative scheduling schemes by using the combined optimization algorithm multiple times to obtain the best logistics scheduling scheme. In each process of using the combined optimization algorithm, when there is a fitness greater than or equal to the fitness threshold Q in the new population, the cycle ends and the algebra corresponding to the new population is obtained; the largest algebra among the multiple algebras is used as the population algebra.

[0087] In this embodiment, it should be specifically explained that the initial population is expressed as: , where A represents the initial population, a r is the rth alternative scheduling scheme, r=1, 2, 3, …, R;

[0088] The fitness function is expressed as: , where f r is the fitness of the rth alternative scheduling solution, COS r is the cost of the rth alternative scheduling plan, including total transportation cost and total personnel cost.

[0089] In this embodiment, it should be specifically explained that step S03 specifically includes:

[0090] Step S031: consider a as a chromosome; a r is the rth chromosome, corresponding to the rth alternative scheduling scheme;

[0091] Step S032: performing natural selection on chromosomes in the population; the natural selection is performed by combining the elite method and the rotation method;

[0092] The elite method generates F1 offspring chromosomes. For a population with a capacity of R, the fitness corresponding to the R chromosomes is arranged from large to small, and each of the F1 chromosomes at the front generates a offspring chromosome; the rotation method generates F2 offspring chromosomes, that is, R chromosomes generate F2 offspring chromosomes according to the corresponding rotation probability; F1+F2=R, so as to keep the offspring population capacity R unchanged and the population increases algebraically;

[0093] The expression of the rotation probability is: , where ζ r is the rotation probability corresponding to the rth chromosome;

[0094] Step S033: performing crossover recombination on the chromosomes in the population; the crossover recombination adopts the PMX method, which is a prior art and will not be described in detail in this embodiment;

[0095] Randomly select E chromosomes from the population for crossover recombination to obtain E new chromosomes; after the chromosome crossover recombination, calculate the fitness of the E new chromosomes, sort the fitness of the E new chromosomes and the fitness of the E chromosomes from large to small and generate a sorting table, and replace the E chromosomes in the population for crossover recombination with the E new chromosomes in the sorting table in positive order; in this embodiment, preferably E=0.7R, if the calculated E is not an integer, round E up to ensure that the calculated E is an integer;

[0096] Step S034: mutate the chromosomes in the population to obtain a new population; the mutation probability is preset to V, and the R chromosomes in the population are mutated according to the mutation probability. The mutation method is to randomly select the positions of two genes in the chromosome and exchange the values ​​of the two genes. In this embodiment, V=0.02 is preferred. The mutation probability is preset by a technician in this field according to the algorithm efficiency and algorithm accuracy.

[0097] In this embodiment, it should be specifically explained that step S04 specifically includes:

[0098] Step S041: The population size is R, each alternative scheduling scheme can be regarded as a bat, and the position of the rth bat is p r , the speed is v r , the sound wave frequency is H r , the loudness of the sound wave is B r , the pulse frequency is I r , bats automatically adjust wavelength and loudness according to the distance between the target and themselves when searching for prey;

[0099] Step S042: Find the optimal position p* of the bats in the current population, and update the position and speed of the bats; the optimal position of the bats in the current population is the bat position corresponding to the alternative scheduling scheme corresponding to the maximum fitness value in this iteration;

[0100] The bat position update formula is: , where p r λ is the position of the rth bat at the λth iteration, p r λ-1 is the position of the rth bat at the λ-1th iteration; v r λ is the speed of the rth bat at the λth iteration;

[0101] The bat speed update formula is: , where v r λ-1 is the speed of the rth bat at the λ-1th iteration, H r is the frequency of the sound wave emitted by the rth bat;

[0102] , where H min is the minimum frequency of the bat's sound waves, H max is the maximum frequency of the sound waves generated by bats, and the sound wave frequency range is [H min , H max ], β is a random vector between [0,1];

[0103] Step S043: Generate a random number rand1 between [0, 1]. If rand1>I r , then random flight is performed, and a new position p is generated near the original optimal position by random flight new , otherwise update the bat position according to the bat position update formula; the formula for random flight is:

[0104] , where p old is the original optimal position, rand2 is a random number between [0, 1], B λ is the average loudness of all bats in the λth iteration;

[0105] When searching for prey, bats will constantly adjust the loudness and frequency of the sound waves they emit according to the location of the target prey to improve their hunting efficiency. As they gradually approach the prey, the spatial range in which the bat searches for prey is gradually decreasing. Therefore, they will gradually reduce the loudness to a fixed value while continuously increasing the frequency in order to quickly and dynamically grasp the location of the target prey.

[0106] Step S044: Generate another random number rand3 on [0, 1]. If rand3 < B r , and the fitness corresponding to the new optimal position is greater than the fitness of the original optimal position, then the position is accepted and the new position is used as the current optimal position, and the sound wave loudness is reduced and the pulse frequency is increased;

[0107] The formula for adjusting the loudness and frequency of sound waves is: ; ;in, , is the sound wave loudness attenuation coefficient, γ>0, is the pulse frequency enhancement coefficient, B r λ+1 is the sound wave loudness of the rth bat at the λ+1th iteration, B r λ is the sound wave loudness of the rth bat at the λth iteration, I r λ+1 is the pulse frequency of the rth bat at the λ+1th iteration, I r 0 is the initial pulse frequency of the rth bat;

[0108] For any sound wave loudness attenuation coefficient and pulse frequency enhancement coefficient, when λ→∞, we have , ; When B r λ When it approaches 0, it can be considered that the bat has found its prey and will not emit a pulse temporarily. The pulse variation range can be set by technical personnel in this field in different numerical intervals according to the specific situation. Only when the bat's position is optimized will the loudness and frequency of the pulse be updated, which suggests that the bat is moving towards the optimal position.

[0109] In this embodiment, it should be specifically explained that the optimal solution in the genetic algorithm at the e-th iteration is used as the starting point of the bat algorithm search at the e+1-th iteration, and the optimal solution found by the bat algorithm at the e+1-th iteration is added to the population of the genetic algorithm at the e+2-th iteration. Therefore, the genetic algorithm and the bat algorithm can share information on the optimal solution, so that both algorithms can benefit from each other's search results, improving the efficiency and accuracy of solving problems, which is more conducive to obtaining the best logistics scheduling plan from complex and changeable scheduling plans.

[0110] In this embodiment, it should be specifically explained that the difference between this embodiment and the prior art lies in that this embodiment has a data classification module, a scheme screening module and a scheme recommendation module. By classifying the data in the real-time logistics data set to obtain the marked data category, the marked data category can be matched with the scheduling content, while the unlabeled data category is not matched, which reduces the amount of data calculation, improves efficiency, and improves accuracy, ensuring that all scheduling contents comply with the latest laws and regulations and policies. A set of alternative scheduling schemes is formed based on the matched scheduling content, which lays the foundation for the subsequent acquisition of the best logistics scheduling scheme. The scheme recommendation module combines the genetic algorithm and the bat algorithm. The genetic algorithm is good at global search, while the bat algorithm is good at local search. The combination of the two can enhance the local search capability missing from the genetic algorithm and the global search capability missing from the bat algorithm. The genetic algorithm and the bat algorithm can share the information of the optimal solution, so that both algorithms can benefit from each other's search results, improve the efficiency and accuracy of problem solving, and thus be more conducive to obtaining the best logistics scheduling scheme from complex and changeable scheduling schemes.

[0111] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0112] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A logistics dispatching management system based on big data, characterized by: It includes data collection module, data classification module, solution screening module, solution recommendation module and intelligent management and control module; The data collection module is used to collect various historical logistics data sets in advance and collect real-time logistics data sets in real time; The data classification module is used to classify various types of historical logistics data sets collected in advance and divide them into N data clusters; and classify the n text data in the real-time logistics data set collected in real time respectively to obtain the marked data category; The scheme initial screening module is used to perform feasibility matching on the original logistics scheduling content and form a set of alternative scheduling schemes; The scheme optimization module is used to obtain the best logistics scheduling scheme from the set of alternative scheduling schemes by combining the optimization algorithm; The intelligent management and control module is used to execute the best logistics scheduling plan and intelligently manage and control the logistics scheduling process; The data categories are the types of logistics data involved in the logistics scheduling process, including vehicles, routes, goods and personnel configuration; The feasibility matching is specifically to match the scheduling content data of the original logistics scheduling solution set with the text content data in the real-time logistics data set, and obtain the alternative scheduling solution set that meets the conditions based on the matching result; The N data clusters are obtained in the following manner: Setting digital labels for the data in the historical logistics data set and the real-time logistics data set; Based on the text data features in the historical logistics data set, the content similarity matrix U is constructed. , where comm(x,y) represents the content similarity between the text data with the digital label x and the text data with the digital label y; the formula is: ； Where k1 is the weight parameter, zx is the feature vector of the text data with the digital label x, and zy is the feature vector of the text data with the digital label y; cov(zx,zy) is the covariance of zx and zy, σ(zx) is the standard deviation of zx, and σ(zy) is the standard deviation of zy; Based on the content similarity matrix, the data in the historical logistics data set are clustered; each text data is regarded as a separate small cluster, and there are Q small clusters in total, where Q is the total number of text data in the historical logistics data set; the gravity value between each two small clusters is calculated based on the content similarity matrix; a merging threshold b is preset to control cluster merging, and the small clusters are iteratively merged. In each iteration, two small clusters q1 and q2 with the largest gravity value are found; If the gravity value between q1 and q2 is greater than the merging threshold b, q1 and q2 are merged into a new cluster q, and the gravity value between the new cluster q and all other small clusters is recalculated; otherwise, no iterative merging is performed; the execution is repeated until the gravity value between no two small clusters is greater than the merging threshold b, at which time N new clusters are obtained, that is, N data clusters; The calculation formula of the gravity value is expressed as: ; Where Y(q1,q2) is the gravitational value between q1 and q2, G is the gravitational constant, c1 is the number of text data in small cluster q1, c2 is the number of text data in small cluster q2, Dis(q1,q2) is the distance between small clusters q1 and q2; that is, the average value of the distance between all text data pairs in the two small clusters; ω is a hyperparameter, σ(q1) is the standard deviation of the feature vector of the text data in small cluster q1, σ(q2) is the standard deviation of the feature vector of the text data in small cluster q2, μ(q1) is the mean of the feature vector of the text data in small cluster q1, and μ(q2) is the mean of the feature vector of the text data in small cluster q2.

2. According to the big data-based logistics scheduling management system of claim 1, it is characterized by: The method for obtaining the labeled data category is as follows: The gravity values ​​of n text data in the real-time logistics data set collected in real time are calculated in sequence with N data clusters, and the gravity values ​​of the i-th text data and the N data clusters are arranged from large to small. The data cluster corresponding to the maximum gravity value is selected as the data category of the i-th text data, and the data categories are marked to obtain m marked data categories. Each marked data category can correspond to text data in multiple real-time logistics data sets; i=1, 2, 3,..., n; and the i-th text data is classified into this data category at the same time.

3. According to the big data-based logistics scheduling management system of claim 2, it is characterized by: The scheduling scheme set consists of multiple scheduling schemes, each data category corresponds to a scheduling content set, the scheduling content set corresponding to the marked data category is all scheduling contents, and the scheduling content set corresponding to the unmarked data category is the scheduling content after the last feasibility match of the category; all the contents in the scheduling content set are combined one by one to form all the scheduling schemes; The original logistics scheduling content is the scheduling content corresponding to the marked data category and the scheduling content corresponding to the unmarked data category. The specific method of forming the alternative scheduling solution set is: Obtain m labeled data categories, perform feasibility matching on the scheduling contents corresponding to the labeled data categories, and obtain a new scheduling content set after matching. The scheduling contents in the new scheduling content set are the scheduling contents after feasibility matching. Feasibility matching includes: Matching the dispatch content corresponding to the marked data category with the text data classified in the corresponding data category; The matching is performed by using keyword matching and semantic analysis or natural language processing related technologies; if a certain scheduling content does not match or conflicts with any text data classified in the corresponding data category, the matching result of the scheduling content is a matching failure; if a certain scheduling content matches all text data classified in the corresponding data category and there is no conflict, the matching result of the scheduling content is a matching success; All successfully matched scheduling contents constitute a new scheduling content set; all new scheduling content sets are combined one by one with the contents of the scheduling content set corresponding to the non-labeled data category to form all alternative scheduling schemes, and the set composed of all alternative scheduling schemes is the alternative scheduling scheme set.

4. The logistics dispatching management system based on big data according to claim 3 is characterized by: The specific method of obtaining the optimal logistics scheduling solution is: Step S01: Encode the alternative scheduling schemes in the alternative scheduling scheme set as a, randomly select R alternative scheduling schemes to construct an initial population, and initialize all parameters; the initial iteration number λ is 0; Step S02: determine the fitness function; when the number of iterations is a plural number, execute step S03; when the number of iterations is a singular number, execute step S04; Step S03: perform multiple iterations, and then directly execute step S05; Step S04: perform a single number iteration, and then directly execute step S05; Step S05: Determine whether the maximum number of iterations has been reached. If so, the iteration ends, and the alternative scheduling scheme corresponding to the maximum fitness value or the optimal position is obtained as the optimal logistics scheduling scheme; Otherwise, let λ=λ+1, and loop from step S02 to step S05; The method of judging whether the maximum number of iterations is reached is: when the fitness is greater than or equal to the fitness threshold Q or converges to a stable value, it is judged that the maximum number of iterations is reached.

5. The logistics dispatching management system based on big data according to claim 4 is characterized by: The initial population is expressed as: , where A represents the initial population, a r is the rth alternative scheduling scheme, r=1, 2, 3, …, R; The fitness function is expressed as: , where f r is the fitness of the rth alternative scheduling solution, COS r is the cost of the rth alternative scheduling solution.

6. A logistics dispatching management system based on big data according to claim 5, characterized in that: Step S03 specifically includes: Step S031: consider a as a chromosome; a r is the rth chromosome, corresponding to the rth alternative scheduling scheme; Step S032: performing natural selection on chromosomes in the population; the natural selection is performed by combining the elite method and the rotation method; Step S033: performing crossover recombination on chromosomes in the population; Step S034: mutate the chromosomes in the population to obtain a new population.

Images