Impact of evolutionary algorithm on optimization of nonconventional machining process parameters

This paper presents the optimization of laser beam machining in additive manufacturing of polymer-based material parameters, specifically focusing on cutting speed, gas pressure of nitrogen, and focal point locations, to achieve optimal mean surface roughness. Using a Python environment, three evolutionary algorithms such as, Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), and Firefly Algorithm (FA), were simulated to evaluate their effectiveness in minimizing surface roughness (Ra). The results of the three algorithms were validated through a benchmark study employing the Genetic Algorithm. The outcomes indicate that the PSO algorithm outperformed the other methods, demonstrating a superior performance in terms of better mean surface roughness. Specifically, the PSO algorithm achieved a mean surface roughness improvement of 0.44% over GA, and 1.1% and 1.23% over ACO and FA, respectively. Notably, the PSO algorithm demonstrated swift convergence, achieving optimal results as early as the second iteration. The PSO algorithm achieved two optimal mean surface roughness values of 0.9333 µm and 0.9838 µm, with an overall average of 0.9399 µm and a standard deviation of 0.0171 µm across 250 runs. These findings indicate that the PSO algorithm excels in delivering superior results while showcasing rapid convergence, robustness, and consistent repeatability in optimizing laser beam machining parameters.