A novel two-phase trigonometric algorithm for solving global optimization problems

Metaheuristics play a major role in the important domain of global optimization. Since they are problem independent, they can be effectively used in constrained and higher-dimension problems and in real-world applications also. In this paper, one novel two-phase trigonometric algorithm is presented to obtain optimal/ near-optimal solutions for different optimization problems. In this work, the main focus is given to solving real-world engineering problems with constraints. The proposed algorithm effectively explores and exploits the search space to arrive at the solutions. Benchmarks analyzed include unconstrained (unimodal and multimodal) functions, constrained special functions, constrained engineering problems and ten problems of the "100 digit challenge", Institute of Electrical and Electronics Engineers (IEEE) Congress on Evolutionary Computation (CEC2019) totaling fifty-nine problem instances. The problems include functions with continuous variables, discrete variables and both continuous and discrete variables. They are all single-objective functions that are suitably modelled and simulated in MATLAB environment. The obtained results are compared with recent and time-tested popular algorithms including Differential Evolution (DE), Improved Teaching–Learning-Based Optimization (ITLBO), Social Network Search (SNS), Firefly Algorithm (FA), Cuckoo Search (CS) and Whale Optimization Algorithm (WOA). Analyses of the results obtained from TP-AB against the performance of the well-known algorithms indicate the superiority and competitiveness of the proposed algorithm in providing quality solutions.