The Research Of Multi-Stage Variable Batch Production Line Reconstruction

As the market demand changes from the mass production of fewer varieties into small quantities of more varieties, assembly line of single species is gradually converted into production line capable of producing a variety of products. However, with the increase in product variety and the fluctuation of the number of single species order caused by changing market demand, problems such as overlarge scale of variable production line, increased number of operatives, long preparation time caused by frequent replacement, low utilization of equipment, low efficiency and poor adaptability of operatives have arisen. Therefore, it imperative to split variable production line into multi-channel production line. Due to the above-mentioned problems extensively exist in manufacturing enterprises of electronic products, conducting research on these problems has important theoretical significance and practical engineering value.Multi-channel is a concentrated variable production line, compared with the variable production line, it has the advantages of reducing production line, less time to prepare for production, easy to product scheduling and so on. In this paper, the input conditions are multi-stage, variable batch, with the smallest batch of similar products to determine a variety of processes, the number of operators and the quantity of equipment in the workplace, the content of the study is the resource allocation and optimization scheduling problem of variable production line splitting under the constraints of equipment resources and human resources, so that a certain period (for example, one year), without changing the structure of the production line under the premise, divide multiple delivery (multi-stage), in each stage, product batch production, batch variable, but there is a minimum batch limit, to minimize the number of tardy.The variable production line split to establish the mathematical model of multi-channel, can be abstracted as nonlinear integer programming problem. Due to a number of issues to be considered after the split the number of production lines, distribution operators, equipment allocation and scheduling optimization, etc. The mathematical model is divided into two layers to solve, the upper layer is resource allocation problem, construction of multi-channel number, operator and equipment in multi-channel distribution plan. The lower layer is product scheduling problem, study the distribution of products with multi-stage, can be divided into batches, variable volume, minimum batch of restricted. The allocation of resources is the precondition of product scheduling, the product scheduling is the result of the allocation of resources. Every problem is a NP-hard problem, the resource allocation and product scheduling method used if loop iteration process, repeated iteration to get relatively optimal solution.This dissertation has carried on the research from following several aspects:First of all, variable production line reconstruct unified model is studied. Aiming at the problem of variable production line reconstruction with the minimum batch limit for the multi-time variable batches with the number of tardiness as the objective function, a unified mathematical model with the objective of minimizing the number of tardiness is established. By analyzing the problems, when there is not much increase the scale of the problem, a sharp rise in the number of decision variables. In order to reduce the size of the solution, resources configuration constraints are increased, the refinement model is given according to the different model.Secondly, the elaboration of multiple channel and product unit model is analyzed. It is proved that the multi-channel resource allocation problem is NP-complete problems, no scheduling rules are given, parallel machine scheduling problem is NP-HARD problem too. The concept of "cell" multi-channel is put forward, and it is used to solve multi-channel resource allocation problems.Thirdly, the improved particle swarm algorithm is proposed, it is used to solve period of change order product scheduling problem at any time. Comparing with other algorithms in the literature, the improved particle swarm algorithm in the objective function and the advantages of running time are reflected. An improved algorithm of optimization scheduling rules is put forward, Giving its lower bound, it is used to solve no period of change orders of product scheduling problem at any time, comparison of traditional optimal scheduling algorithm with non-batch EDD +FCFS, EDD+LPT, EDD+SPT, EDD+RWD.Fourth, the improved AFSA(artificial fish-swarm algorithm)+Improved particle swarm algorithm is used to solve the decision problem of large-scale multi-channel segment changes at any time order the joint; The improved AFSA+DM-LPT scheduling algorithms is used to solve decision problem of rule changes at any time not of the order of large-scale multi-channel joint; Enumeration method+improved particle swarm algorithm is used to solve small-scale multi-channel, the joint decision-making problem period of change orders at any time, enumeration method+DM-LPT rule scheduling algorithm is used to solve not subject to change orders segment of small-scale multi-channel joint decision problem.The innovation of this dissertation is the reconstruction of production line with multi-time dynamic scheduling with the objective of tardiness as the objective function.