Genetic algorithm-based integrated approach to flexible manufacturing system part type selection and machine-loading problem

Part type selection and machine loading are two interrelated subproblems in production planning of flexible manufacturing systems. The total solution requires a simultaneously combined approach to avoid the possible conflicts between the two sets of individually obtained solutions. A strict mixed-integer programming (MIP) model that integrates part type selection and machine loading together is formulated. The MIP takes into account the constraints such as magazine capacity, tool life, available machine time, etc. The objective is to minimize the difference between maximum and minimum workloads of all the machine resources in each batch. A genetic algorithm-based method is developed to obtain the solution of the problem effectively. Concepts of virtual job and virtual operation are introduced in the encoding scheme, and a chromosome is composed of both these strings. Among each chromosome, the partition symbol list is mainly used to handle the part type selection problem, while the virtual job list is mainly used to cope with the loading problem. Special crossover and mutation operators are designed to adapt to the problem. Our approach can simultaneously balance the workloads in different batches. At last, illustrative examples are presented, and a comparison between standard MIP algorithm and a genetic algorithm method is given.