Dynamic Multi-intelligent Algorithm And Its Application In Logistics And Distribution System

With the rapid development of economy and population,the industrial agglomeration effect has been further deepened;the level of user consumption and purchases have continuously increased,and logistics activities have increased dramatically.Therefore,logistics vehicles will inevitably increase.The emission and energy consumption of these logistics vehicles are far greater than the passenger vehicles used for daily travel.Undeniably,the efficient and efficient modern logistics industry plays a key role in promoting economic development.In addition,it also caused a series of problems such as energy consumption,environmental pollution and traffic problems.How to reduce the negative impact of logistics activities from the logistics and distribution system has become one of the important issues in the modern logistics industry.The reason is that the industry has not yet formed a complementary and win-win situation,and the distribution is simple and monotonous.The functions of logistics enterprises cannot meet the needs of diversification and multi-level logistics and distribution.A survey from the investment adviser shows that less than half of the logistics enterprises to establish a logistics information management system and internal network resources.Among them,80% of the logistics enterprises are still in the initial stage of application.About 15% of the logistics enterprises are in the process of transformation and control of the optimization stage.Only 5% of the logistics enterprises have been raised to the supply chain level.The development of cloud computing is an important transition and symbolic change in the IT field.Cloud is an effective and usable virtualized asset and is an interconnected network of various servers sharing assets.Cloud computing becomes efficient through virtualization.Therefore,it becomes more important to effectively manage and allocate virtual machines in a cloud environment.In this article,we analyze and improve the reliability of cloud computing infrastructure while allocating resources.The introduction of cloud computing into the logistics industry can solve the problems of insufficient development and utilization of information resources and the lack of information resources in the industry,and at the same time promote the transformation of China's traditional logistics industry into smart logistics;introducing cloud computing into logistics distribution,even to a greater extent.The field of distribution services is an urgent problem for the logistics industry.Logistics distribution is not only a key link in the logistics service supply chain,but also an important part of e-commerce activities.Importantly,the study of vehicle scheduling problem is conducive to the development of intelligent transportation,is conducive to building a comprehensive logistics and distribution system,more conducive to the development of e-commerce.Therefore,the problem of vehicle scheduling should be more extensive in real life,such as vehicle routing problems,traveler problems,postal delivery problems,flight scheduling,terminal ship handling and power system scheduling problems.Not only that,the development of e-commerce and communication technology is to promote the combination of vehicle scheduling and logistics and distribution,so that in the chain of supermarkets,shopping malls and express delivery industry and other areas with a more comprehensive development platform and a broader application prospects.The dissertation main research content and innovation are as follows:(1)In this dissertation,the theoretical basis of rough set-based generalized decision-making information system is discussed preliminarily.We redefine the concept of discernibility matrix,and regulate Pawlak reduction,generalized decision reduction and optimal decision rules.We focus on the logistics and distribution system based on cloud computing service platform.We need to analyze the customer's real-time needs and pre-process the massive demand data.(2)Starting with solving the convergence and solution set distribution of complex multi-objective optimization problems,this dissertation proposes a multi-objective differential evolution algorithm based on population adaptive adjustment.First,we design a strategy of population expansion,and generate some new individuals in the decision space to search for better non-dominated solutions.Secondly,we design a strategy of population contraction,and eliminate the poor individuals according to their contribution to the non-dominated solution sets to reduce Calculate the load and reserve some space for new disturbance individuals with population diversity.Then,this dissertation introduces the elite learning strategy to prevent the algorithm from falling into local convergence.Finally,using 6 sets of function optimization problem verification,in most cases the new algorithm is obviously superior to the MOPSO and MODE algorithms in terms of convergence and diversity,and its performance is analyzed and evaluated.A TSP test instance Gr17 verifies that the new algorithm can effectively solve the MTSP problem.Therefore,the multi-objective differential evolution algorithm based on population adaptive adjustment can well balance global search and local search and can be used as an efficient algorithm for resource allocation.(3)This dissertation analyzes the limitations of GA and PSO,and then introduces crossover operators and mutation operators in genetic algorithms to design a new PSO algorithm with genetic operators.Firstly,the particle swarm is initialized randomly.After the fitness value of each particle is calculated,the local optimal solution and the global optimal solution are dynamically updated.Then the cross-over position is selected with a certain probability,and the particles are selected circularly from the population.In turn,the local optimum and the global optimal crossover of the same population of particles are intersected.Finally,the optimal solution is obtained by the variation of the particle itself.Through the test cases of five different cities,the path cost of the TSP problem is simulated,and the optimal convergence state and convergence times are achieved.Compared with the PSO algorithm,the feasibility and validity of the improved genetic-particle swarm hybrid algorithm.(4)Based on a detailed analysis of vehicle cost,this dissertation uses the advantages of cloud computing and rough set to get the customer demand and vehicle load in the distribution network;and proposes a genetic algorithm based on cloud computing.Firstly,in the process of crossover and mutation,adaptive crossover probability and mutation probability are generated by the cloud model condition generator algorithm.Then,the algorithm performs crossover and mutation operation and analyzes the performance parameters.The simulation results show that the cloud computing method is effective for the optimization of vehicle routing in cold chain logistics,which can provide real-time optimization path within a certain period and control the cost of cold chain logistics and distribution to improve the efficiency of delivery service.The solution strategy can Timely response to dynamic demand,well meet the requirements of real-time dynamic vehicle routing problem.(5)To effectively improve the service performance of cloud computing and improve the utilization efficiency of resources,this dissertation introduces the credibility evaluation mechanism,the local pheromone updating mechanism and the global pheromone updating mechanism and proposes a cloud computing based ant colony algorithm.Compared with the optimal time scheduling algorithm and the standard ant colony algorithm,the experiment proves the correctness and effectiveness of the improved ant colony algorithm in cloud computing resource scheduling.Compared with other resource scheduling algorithms,ACO based on cloud computing not only has obvious advantages in single-target and multi-task resource scheduling,but also has a good realization of the actual effect of cloud computing resource scheduling application,which helps users quickly find the optimal virtual machine node and allocate resources to the logistics center.