An improved bat algorithms for optimization design of truss structures

Abstract

This paper exploits a novel improved metaheuristic to entrance the standard bat algorithm (BA) performance for optimization engineering problems and optimization design of truss structures. As is the case with many metaheuristic algorithms, BA is straightforward to implement, and such a basic algorithm may address a wide variety of problems with great flexibility. However, when its exploitation stage is changed too quickly by rapidly shifting the loudness and pulse emission rate, it may stagnate after an early stage. It could result in a low probability of success in finding the real prey. Therefore, a new loudness function is proposed to adapt the step size of the random walk efficiently. Also, this study proposes a push-process technique that can generates an intervention in the BA algorithm to expose the true global optimal in a few generations instead of many calculation times. Then, all results are validated to those of well-established research. Finally, the proposed BA is applied to optimize the structural weight of truss structures.