Principle and performance validation of search and rescue team algorithm

Combinatorial optimization problems exist in various fields. However, their solution is uncertain, particularly for high-dimensional non-deterministic polynomial-time hard (NP-hard) problems. To achieve a more stable solution for high-dimensional problems, a new metaheuristics algorithm, called the search and rescue team (SaRT) algorithm, is proposed herein. This article presents its mathematical model and verification process using fifteen test case studies including five knapsack problems and ten multi-modal test functions, also taking optimization of the crank slider mechanism as an engineering application case. The parameter sensitivity of the SaRT was studied and its output was found to be insensitive to the parameters. The SaRT can converge to the global optimum with a probability of more than 96 % for NP-hard problems, and also shows excellent performance and robustness in multi-modal test functions and multi-objective optimization compared with other popular algorithms. Hence, the proposed algorithm can be employed for optimization design and parameter tuning in engineering control and fault diagnosis.