A Hybrid Smart Quantum Particle Swarm Optimization for Multimodal Electromagnetic Design Problems.
Shah Fahad,Shiyou Yang,Shafi Ullah Khan,Shoaib Ahmed Khan,Rehan Ali Khan
DOI: https://doi.org/10.1109/access.2022.3188276
IF: 3.9
2022-01-01
IEEE Access
Abstract:Quantum behaved particle swarm optimization (QPSO) has been one of the most widely used algorithm in engineering world. Since its debut in 2004, QPSO has been used for resolving numerous complicated multimodal problems. Moreover, considering the adaptability and versatility, it has resolved a variety of real-world and test problems. To tackle numerical and engineering optimization problems, we introduce novel hybrid algorithm QPSODE. The novel hybrid algorithm integrates Quantum behaved particle swarm optimization (QPSO) with differential evolution (DE) strategy. A crossover and selection (influenced by DE) is used in the QPSODE’s position updating mechanism. During the selection process, the Boltzmann operator is applied to the position vectors of two randomly chosen particles, not to their individual optimum placements. Therefore, unlike the QPSO, a particle is only relocated to a new position if it has a higher fitness value, implying the application of a selection strategy across the whole search space. Additionally, the hybrid algorithm is improved by introducing proper parameters tuning, control parameter, path disparity. The hybrid algorithm enhances the algorithm’s performance by speeding up the convergence and avoiding the premature convergence, the main flaw in the earlier algorithms. The proposed algorithm is put to test, by using 19 well-known benchmark test functions and the engineering optimization problem for superconducting magnetic energy storage (SMES). In terms of the quality of the resulting outputs, QPSODE outperforms various state-of-the-art approaches.