Surface Roughness Prediction and Optimisation using Novel Joint Artificial Neural Network and Bat Algorithm

This paper targets the surface roughness concept in end milling in which the tool-work material combination is central to its success. At present, sufficient optimal surface roughness information is repeatedly not accessible to CNC end milling operators and this problem is anticipated to grow worse in the forthcoming years. Consequently, the unique development and validation of optimisation tools are interventions to tackle access to optimal roughness information problems. This paper examined two novel models, the combined artificial neural network and bat algorithm as well as joint artificial neural network and particle swarm optimisation to predict and optimise the process parameters of an end milling scheme. Both models were tested with literature data. Additionally, the work investigates machining time and introduces a bi-objective fuzzy goal programming optimisation model. The striking results revealed the optimal values as 0.8816 and 0.8088 for the particle swarm optimisation procedure while the bat procedure yielded 0.275 and 0.178, which places the bat procedure ahead of the counterpart, particle swarm optimization procedure