DTCSMO: An efficient hybrid starling murmuration optimizer for engineering applications

Abstract

Starling murmuration optimizer is a newly well-developed swarm intelligence algorithm inspired by the behavior of starlings during stunning murmuration and has performed competitively with other well-known methods. However, starling murmuration optimizer faces the issues of poor search capability, iterative stagnation, and low convergence accuracy when dealing with complex engineering optimization and high-dimensional problems. To ameliorate these deficiencies, an efficient hybrid starling murmuration optimizer that combines dynamic opposition, Taylor-based optimal neighborhood strategy, and crossover operator is developed in this paper. Firstly, introducing the Taylor-based optimal neighborhood strategy ensures a more careful and reasonable reliance on optimal individuals, thus reducing the possibility of falling into local optima. Meanwhile, dynamic opposition successfully balances the exploration and development phases by incorporating the ideas of quasi- and reflective opposition. In addition, the exploration-development capability is further enhanced by introducing the crossover operator. To ensure the adaptivity of the crossover factor, 12 different inertia weights are explored for crossover factors, and the sigmoidal decreasing inertia weights achieve optimality. Finally, the excellent performance of the proposed algorithm is verified by comparing it with the improved algorithm and the state-of-the-art search algorithms on 26 benchmark functions and the CEC2020 test suite, respectively. In addition, the proposed algorithm is used to optimize ten practical engineering optimization problems. Experimental results demonstrate that the proposed algorithm has strong competitiveness in this algorithm regarding computational power and convergence accuracy. Finally, the proposed algorithm is applied to two truss topology optimization designs. The experimental results demonstrate the applicability and potential of the proposed algorithm in practical applications. The proposed algorithm is a competitive optimization algorithm for solving optimization problems.