Development and application of Quantum Entanglement inspired Particle Swarm Optimization

Abstract Particle Swarm Optimization has been extensively researched and applied to tackle optimization problems due to the ease in implementation and less number of parameters to be tuned. But particle swarm optimization (PSO) algorithm gets trapped into local optimum in high-dimensional space and it is inefficient in solving optimization problems which show high dependency. To overcome the above problems without compromising the advantages of PSO, this paper proposes Quantum Entanglement inspired Particle Swarm Optimization (QEPSO). QEPSO incorporates entangled states in its Q-bits to efficiently solve high-dependency problems and uses quantum local search to accelerate the optimization process. The proposed algorithm is tested on several standard benchmark functions and is also further benchmarked on IEEE Congress of Evolutionary computing (CEC 2017) benchmark set. The performance of QEPSO is compared with existing variants of PSO and some other popular algorithms. The results show that QEPSO outperforms other algorithms and is especially useful in high dimensional problems. Finally it is used for a real-life application of Multi-level Image Segmentation where eight gray-scale standard test images were used. The performance of QEPSO was superior to the other algorithms as it gave better results with high stability and quick convergence.