Simultaneous assembly planning and assembly system design using multi-objective genetic algorithms

This paper aims to demonstrate the application of multi-objective evolutionary optimization, namely an adaptation of NSGA-II, to simultaneously optimize the assembly sequence plan as well as selection of the type and number of assembly stations for a production shoo that produces three different models of wind propelled: ventilators. The decision variables, which are the assembly sequences of each product and the machine selection at each assembly station, are encoded in a manner that allows efficient implementation of a repair operator to maintain the feasibility of the offspring. Test runs are conducted for the sample assembly system using a crossover operator tailored for the proposed encoding and some conventional crossover schemes. The results show overall good performance for all schemes with the best performance achieved by: the tailored crossover, which illustrates the applicability of multi-objective GA's. The presented framework proposed is generic to be applicable to other products and assembly systems.