Research And Development Of Mixed Model Assembly Line Balancing System

As current markets are characterized by a growing trend for higher product variability, mixed-model assembly lines (MMAL) are preferred over the traditional single-model assembly lines. For MMAL, it is desirable that the line be balanced. Balancing the assembly line means that workstations in the line are running at the same pace without any idle time.The main outcomes of this thesis work are, firstly, mathematical models are used to discuss the mixed-model assembly line balancing problem (MMALBP). Various objectives and various types of mixed model assembly line balancing problems are modeled. These models overcome the limitations of traditional mathematical models and also provide some good methodology and tools for production managers.Secondly, a design of genetic algorithms as solutions to various types of MMALBP is proposed in this thesis. In terms of solving MMALBP, genetic algorithms are designed for the different mathematical models. The genetic algorithms are modified on three places: the production of initial population is studied; a unique, practical and simple principle for gene encoding and decoding is developed; genetic arithmetic operations are developed for mixed model assembly line balancing.Finally, a design of a mixed model assembly line balancing system is presented. In addition, precedence diagram modeling is implemented that models the constraints of the tasks in a mixed model assembly line so that the user can describe and modify the model easily. The system calls algorithms which are compiled to libraries to obtain balancing results. Where a solution is found, it can be displayed graphically. The system is tested using practical project data.