A modified algorithm of Particle Swarm Optimization for form error evaluation

The set of measured data points acquired from the Coordinate Measuring Machine (CMM) need to be processed and analyzed for evaluating the form errors inside the manufactured components. This paper presents a modified algorithm of particle swarm optimization (MPSO) for assessing the form error from the set of coordinate measured data points. In the classical algorithm of the particle swarm optimization (PSO), the value of the candidate solution is updated from its existing value without actually comparing the value obtained in the consecutive iterations for fitness. This behaviour attributes to a lack of exploitation ability in the defined search space. The proposed algorithm generates new swarm position and fitness solution for the objective function through an improved and modified search equation based on a proposed heuristic step. In this step, the swarm searches around the best solution of the previous iteration for improving the swarm exploitation capability. The particle swarm uses greedy selection procedure to choose the best candidate solution. A non-linear minimum zone objective function is formulated mathematically for different types of form errors and then optimized using proposed MPSO. Five benchmark functions are used to prove the effectiveness of the modified algorithm, which is verified by comparing its solution and convergence with those obtained from the established algorithms namely PSO and genetic algorithm (GA). Finally, the result of the proposed algorithm for form error evaluation is compared with previous work and other established nature-inspired algorithms. The results demonstrate that the proposed MPSO algorithm is more efficient and accurate than the other conventional heuristic optimization algorithms. Furthermore, it is well suited for form error evaluation using CMM acquired data.