Research On Stochastic Mixed-Model Assembly Line Balancing Problem Of Type Ⅱ

Assembly line is a perfect interpretation of industrial engineering techniques, the assembly line producing fully embodies the industrial engineering ideas, such as" specialization of working","standardization","manufacturing technique proceduring". It can improve production efficiency and reduce cost, which mainly reflects the advantages of "quantity", but the customer demand-oriented is becoming more and more obvious and the flexibility also becoming more and more important. The traditional simple assembly line can’t meet requirements rapidly. So balancing of stochastic mixed model two-sided assembly line is studied in the thesis.Compared with the simple assembly line, stochastic mixed model two-sided assembly line possess much more advantages as follows:Firstly stochastic is more realistic, because there is always a wide variety of random factors in practical production; Secondly, the "mixed" is more adaptive to market; Finally, The "two-sided" is necessary for the production of many large machines and equipments, and it can shorten the length of the assembly line, improve the usage rate of equipments and labor.It takes Y automobile manufacturing company’C line as the research object in this thesis and seeks the optimal solution to type Ⅱ balancing problem of stochastic mixed model two-sided assembly line. The process, methods, the main work done and major achievements obtained are as follows:(1) It is classified random factors of assembly line into five categories:"man","machine","material","environment" and "method". Impact of randomness on assembly line balancing is converted into the impact on the time process by combining work sampling and statistical analysis. According to statistics, the process time appears in the form of a distribution. Distributions which is obtained by work sampling and statistic may be different as different subjects. The process time of this thesis follows normal distribution.(2) The mathematical model of type II balancing problem of stochastic mixed model two-sided assembly line is established. The mathematical model considers the random factors of assembly line, impact of mixed model and wait time of two-sided assembly line, converts the impact of randomness to assembly line balancing into the impact to the process time, and computes the comprehensive process time.(3) Genetic algorithm based on a feasible sequence real-coded is designed. Firstly to design a suitable and effective fitness operator; then to find the optimum relationship of crossover rate and mutation rate that genetic algorithm can reach global convergence quickly and effectively; finally to solve the classical assembly line balancing problem through different algorithm, the compare of the results proves that the genetic algorithm is effective.(4) The production cycle of C line in Y automotive company is optimized. The cycles before and after optimizing are144s and104s respectively, which has been decreased27.78%. So it demonstrates the effectiveness of the genetic algorithm.The conclusions are as follows:Firstly handling stochastic factors by combing work sampling with statistical methods is effective; Secondly the designed genetic algorithm in this thesis is an efficient algorithm to solve stochastic mixed model two-sided assembly line balancing problem of type Ⅱ; Finally the research is innovative to a certain degree and provides theoretical reference for solve stochastic mixed model two-sided assembly line balancing problem of type Ⅱ.