Research On Some Buffer Capacity Allocation, AGV Path Planning And Other Production Issues

With the increasing development of science and technology,the core problems of many projects ultimately come down to optimization problems.Pursuing the overall optimization of the system based on a system perspective,and carrying out optimization design,resource allocation,scheduling,planning,control and management,etc.have important functions and significance for the development of modern enterprises,industrial production and related technologies.The optimization of production systems affected by random events has been the focus and diff iculty of research in this field due to the complexity of the problem and the limitation of the solution technology.In this paper,by the national natural fund support,consider including unknown system information,the influence of random events industri al production practice,some production site problems such as optimal buffer capacity allocation and AGV distribution path planning are studied.The main work carried out as follows:1.Aiming at the problem of optimizing buffer capacity allocation of non-similar hybrid production lines affected by random events such as equipment failure,a hybrid fuzzy clustering buffer capacity allocation technology based on genetic algorithm was proposed.This technology combines with Gershwin decomposition evaluation technology,introduces the fuzzy clustering idea on the basis of the genetic algorithm,and carries out fuzzy clustering on the equipment according to its parameters.Different from the traditional method,which determines the initial solution according to the mean partition method,the initial solution is determined according to the similarity degree of adjacent equipment.At the same time,a new population updating strategy is proposed in the process of optimization by combining the non-similar conversion technology of hybrid production lines with the characteristics of buffer capacity allocation problem.In particular,a fast evaluation function is introduced to evaluate and control the new solution and optimization direction,which improves the optimization efficiency.The simulation results show that this technique has better performance than the traditional genetic search technique.2.The optimal allocation of buffer capacity in scale hyb rid production lines has not been effectively solved in this research area.Aiming at this problem,a hierarchical buffer optimization allocation technology based on improved simulated annealing algorithm is proposed.This technology combines the traditional hybrid production line decomposition and comprehensive technical ideas,and proposes a hierarchical analysis method for scale hybrid production lines.The scale production line system is decomposed into several production line subsystems containing virt ual production equipment.The optimal solution of the subsystem is solved to construct the progressive optimal solution of the original system,which greatly improves the efficiency of optimization.At the same time,in the optimization of the subsystem,combined with the Gershwin decomposition evaluation technology,the crossover operator and the reversal operator are introduced,and an improved simulated annealing algorithm is proposed.The crossover operator can improve the gene renewal efficiency and speed up the optimization,while the reversal operator can give a disturbance to the solution of each generation with a certain probability to make it jump out of the local optimal solution and increase the global search ability.Simulation experiments show that this technique not only performs well for general systems,but is especially suitable for solving the problem of buffer capacity optimization of scale hybrid production lines.3.Aiming at the problem of path planning of AGV with unknown obstacles in material distribution of production line or buffer zone,a new directed D* Path Dynamic Planning Algorithm is proposed.This algorithm introduces the concept of key nodes in the distribution environment,determines the feasible path of AGV in a step-by-step expansion way,and introduces a guiding function to reduce the searching space of path planning,and improves the searching efficiency.At the same time,a smoothness function is introduced on the basis of the traditional Euclidean distance evaluation standard,which penalizes the deviation degree between the planned path and the ideal path,reduces the number of inflection points,and improves the smoothness of the path.In this paper,the method of determining the relevant turning factors in smoothness function and their characteristics are described in detail.Based on this,the convergence of the directed D* algorithm is proved and the algorithm flow is given.Simulation experiments show that the algorithm has higher planning efficiency than the traditional AGV path dynamic planning algorithm.4.Finally,combined with the GUI of MATLAB,an algorithm tool software module for the above problems was developed and tested experimentally.