LAUSR

Harmony search algorithm (HSA), inspired by the behaviors of music improvisation process, is a widely used metaheuristic to solve global optimization problems arises in various fields. The reason of its popularity is its simplicity of algorithm structure and good performance. However, the conventional harmony search algorithm (HSA) experiences prone towards the local optima, tedious task of tuning parameters, and premature convergence. To overcome all these drawbacks of conventional HSA and further improve the precision of numerical results, a new variant of the HSA called modified harmony search algorithm (MHSA) is proposed in the present study. This MHSA utilizes the valuable information stored in the harmony memory and modifies the search strategy to make an efficient search procedure by adopting a new formulation to the pitch adjustment process, randomization process, harmony memory considering rate (HMCR), and pitch adjustment rate (PAR). The experimental validation and comparative performance study with conventional HSA, variants of HSA such as adaptive harmony search with best based search strategy (ABHS), enhanced self-adaptive global-best harmony search (ESGHS), novel self-adaptive harmony search (NSHS), parameter adaptive harmony search (PAHS), Gaussian global-best harmony search algorithm (GGHS) and other metaheuristics such as sine cosine algorithm (SCA), grey wolf optimizer (GWO), comprehensive learning particle swarm optimization (CLPSO), gbest-guided artificial bee colony (GABC), and covariance matrix adaptation evolution strategy (CMA-ES) is conducted on a set of 23 well-known benchmark problems. In addition to this benchmarking, the proposed MHSA is also used to solve three structural engineering design problem. The statistical test and convergence behaviour analysis are used to analyze the quality of search and significance of improved accuracy. The comparison illustrates the superior search efficiency of the proposed MHSA than other algorithms as a global optimizer.