Research On Storage Location Assignment Problem Based On Improved Artificial Fish Swarm Algorithm

The living style of people is changed due to the development of smart devices and network technology.Shopping online has become one of the important ways of purchase.On one hand,with the acceleration of the pace of life,consumers have higher and higher requirements for the delivery time of goods in shopping online.On the other hand,due to the very frequency of ecommerce warehouse picking operations,the time consumed has accounted for more than half of the order response time.Therefore,the work of optimizing storage location assignment(SLA),reducing picking time and improving order response speed have played an important role in warehouse management of e-commerce enterprises.The SLA problem is studied,that include three main contents: the improvement of intelligent optimization algorithm,the establishment of the SLA model and the application of algorithm under one specific condition.The first research focus is the improvement of artificial fish swarm algorithm(AFSA).The optimization of SLA belongs NP-Hard problem,and there is no effective algorithm in polynomial time.So scholars mostly use intelligent optimization algorithm to solve it.The demerit of the current improved AFSAs are analyzed,and it is considered that there is still room to improve the basic behavior efficiency of the fish swarm.According to the interaction between individuals and the population,the basic behaviors are divided into social behaviors and individual behaviors,and different visual fields and state change speeds are set for the two types of behaviors.On the one hand,it strengthens the interactivity of social behaviors,so that the fish swarm can quickly approach the found extreme value;on the other hand,it enables individual behavior to search the surroundings in a more fine-grained manner to find an accurate and better solution.In addition,in order to solve the problem that the variance of the results of multiple calculations is large when solving high-dimensional problems,some ideas from genetic algorithms and bionics are absorbed,then three new behaviors of Gaussian Mutation,Dimension Optimization and Chaotic Mutation are introduced.The ablation experiments and comparison experiments show that the improved algorithm gets the more accuracy of the solution results,and the less difference of multiple calculations with the stronger robustness.Secondly,modeling SLA problem is another research focus.Starting from the enterprise goal,the strategy of SLA is analyzed,and the existing model is not fully applicable to these strategies.According to the characteristics of the research problem,a new SLA model is established,and three optimization sub-goals are proposed for the e-commerce warehouse,namely,shortening the time for goods in and out of the warehouse,the convenience of nearby storage of related goods,and the convenience of picking up goods.Then the coefficients are assigned to each sub-goals by weight distribution,and to minimize the weight sum is the final optimization goal.Finally,the adaptation and application of the improved AFSA in SLA are studied.First,starting from the coding rules of artificial fish,the dimensions of the problem is reduced to onethird of the original through dimensionality reduction mapping,which reduces the computational difficulty.Second,the original Euclidean step of the fish swarm is changed to a dimensional step,and the discretization process is carried out,so that the algorithm can be adapted to discrete problems.Third,the initial storage is optimized by Opposition-based Mutation,and the iterative effect is improved by Dimension Exchange,so that the fish swarm algorithm for SLA is formed,which is named SLA＿AFSA.It is shown in the comparison experiments that the optimization effect of SLA＿AFSA is improved by more than 2.8%.Numerical analysis and the output storage location indicate that the three sub-objectives of the model are effectively optimized.