Abstract: In view of the disadvantage of traditional production mode and the requirement of lean production, a standpoint is put forward on the basis of analyzing lean production system operation. The point is that lean production operation optimization is defined as large-scale system multiobjective programming problem with angular structure of the original box. An operation optimization model is constructed by taking workload of manufacturing factors in plant as decision-making variables and the indexes of output, quality, cost and delivery as dynamic constraint conditions. In the model, priority and its factor are introduced into objective function for reflecting important degree of multiple mutually conflicting objectives. So the best operation objective requirement is determined on the basis of dynamic attribute of actual problem. The characteristic of the model is dynamic, multi-objective and multistage. The algorithm called “harmonizing cluster results” for the model is presented in order to solve the difficulties of solution for the lean production system operation model. Then the model is simplified into some submodels, such as production ability optimization submodel, time optimization submodel, balancing and harmonizing submodel. The method of “harmonizing cluster results” is used to solve the above submodels. Finally the lean production operation optimization model and its solution algorithm are applied to an assembly line of energy meter in Zhejiang Province. The results of Witness simulation verify that the model is rational and its solution algorithm is accurate and highly efficient.

Key words: Harmonizing cluster results, Lean production system, Modeling simulation, Multi-objective optimization, Operation optimization model