Performance Analysis of Gravitational Search Algorithm During Parametric Optimization of Machining Processes

Optimization of machining parameters is usually carried out to enhance the performance of the machining processes. Generally, the design of experiments approach and multi-response optimization methods are applied to obtain the best parametric setting of a process. However, these methods result in local optimal solutions because the search is limited to discrete points. Therefore, classical optimization methods and metaheuristics are extensively used to obtain the optimized machining condition. Cuckoo search (CS), teaching learning-based optimization (TLBO), particle swarm optimization (PSO), and genetic algorithm (GA) algorithms have proved their efficiency for optimizing machining processes. Despite outperforming evolutionary and swarm intelligence techniques on unconstrained benchmark functions, few studies have used a gravitational search algorithm (GSA) to obtain the best parametric condition in a machining process. Therefore, the present study attempts to quantify the performance of GSA and chaotic GSA (CGSA) while determining the best parameters for a machining process. The present study adopts different cost functions involved in turning, wire electrical discharge machining (WEDM), and plasma-enhanced chemical vapor deposition (PECVD). A comparison of results obtained with TLBO and PSO indicates the superiority of TLBO by maintaining better mean and standard deviation values. However, CGSA performs significantly better than GSA and PSO. Friedman’s test ranks TLBO as the best algorithm, followed by CGSA, PSO, and GSA.