A modified cuckoo search algorithm to optimize Wire-EDM process while machining Inconel-690

Abstract

Cuckoo search (CS) algorithm was found to be efficient in yielding the global optimal value, and this algorithm was found to outperform genetic algorithm (GA) and particle swarm optimization (PSO) techniques. However, the accuracy of CS heavily depends upon the initial solution and its location from the target value and, therefore, it may involve many generations. Furthermore, the evolutionary operators are applied in each generation. This could lead to delay in convergence. To improve the performance of cuckoo search further, an attempt has been made in the present work to propose a modified cuckoo search involving two-stage initialization. Benchmark functions have been used to test the performance of the proposed method. Furthermore, the proposed method has been applied to wire electrical discharge machining (WEDM) process. Inconel-690, a nickel-based superalloy, has extensive applications in aerospace and nuclear power sectors. Although WEDM is one of the advanced machining processes used to machine such hard-to-cut materials, machining data for this material is not available in the literature. The proposed algorithm was found to be accurate and fast as compared to the GA, PSO, and existing cuckoo search. The machining data generated in this work will also be useful to the industry.