Research On Multi-objective Mixed-model Assembly Line Balancing Problem Based On IWD Algorithm

Mixed-model assembly line balancing problem is a key technological problem which the manufacturing must face. Solution to the problem directly affects the production efficiency and the cost of product. Through reasonable improvement, on one hand, the equipment utilization of enterprise can be maximized, on the other hand, the waste of resources can be reduced, thus realizing lean manufacturing. However, due to the large variety of products, the variable proportion of each type, the differences of tasks and task time, mixed-model assembly line balancing problem is more complex than that of single-model. Therefore, conducting research on mixed-model assembly line balancing problem is of theoretical significance and economic value.In this paper, according to the characteristics of mixed-model assembly line balancing problem, multiple evaluation function that is close to the actual production process is established with considering the influence factors of production. In addition, based on IWD algorithm, solving method to multi-objective mixed-model assembly line balancing problem is put forward. The main work done is as follows:(1) IWD algorithm is introduced into mixed-model assembly line balancing problem areas: First, the encoding rules, local soil update rules and global update rules are improved. Then, in order to increase the global optimization ability of the algorithm, the node selection rule is redefined by joining the largest probability selection rule and random search rule, thus forming hybrid selection mechanism. Finally, the effectiveness of the improved IWD algorithm is verified through several standard problems. What’s more, compared with genetic algorithm, experimental results show that solving rate and the effect of IWD algorithm is superior.(2) Mathematical model of multi-objective mixed-model assembly line balancing problem is established by integrating maximal task relatedness, minimum workstation number and minimum workload balance coefficient, so that the model is consistent with the actual production process. For multi-objective mixed-model assembly line balancing problem, the method of Pareto dominant sorting is used to evaluate multiple targets. Then on the basis of the sorting results IWD algorithm was improved to increase the global optimization ability. Finally, 10 standard problems are used for test and the test results show that task allocation scheme is more consistent with the actual production when considering multi-objective function.(3) By an example of a firm’s LR series engine assembly line balancing problem, the effectiveness of the improved IWD algorithm and the rationality of the mathematical model are verified.