SUMMARY

Hierarchically coupled constrained optimization problems (HCCOPs) are commonly encountered in the manufacturing industries, however, they haven’t been categorized as such. Due to a lack of clear definition to identify these problems, numerous techniques have been developed for the optimization of HCCOPs but these techniques are not universally applicable to all HCCOPs and are unable to cope with large scale problems. Furthermore, current techniques used for the optimization of these HCOOPs only provide a single optimal solution upon execution. Though the single optimal solution may theoretically satisfy the objective function, it might not be applicable in real life scenario. This research will first focus on establishing an abstract definition and identifying the common principles amongst different HCCOPs. Next, based on the established definition and common principles, a new optimization technique, based on evolutionary computation, will be developed. The proposed algorithm will be developed in a way such that only feasible solutions are generated during the iterations of the algorithm thereby reducing its computational complexity. To validate the proposed algorithm, it will be utilized to optimize HCCOPs such as assembly job-shop scheduling problem (AJSSP) and the simultaneous optimization of Neural Network (NN) structure and weight values. The research will also focus on developing a technique for the multimodal optimization (MMO) of HCCOPs i.e. obtain multiple solutions with the same objective value. To validate the proposed MMO approach, it will first be utilized for the MMO of benchmark job-shop scheduling problem (JSSP) and permutation flow-shop scheduling problem (PFSSP) followed by the MMO of AJSSP.