Tenrec Optimization Algorithm: A Novel Swarm-based Metaheuristic Algorithm Inspired by the Natural Foraging and Digging Behaviors of Tenrec

Metaheuristic algorithms are important optimization tools in engineering sciences and real-world applications due to their high ability to solve complex, nonlinear, and multidimensional problems.The success of these algorithms largely depends on striking a proper balance between exploration and exploitation during the search process, because exploration leads to extensive exploration of the search space and avoids getting stuck in local optima, while exploitation accelerates the convergence process towards high-quality answers by focusing on promising areas.Accordingly, providing new strategies for effectively managing these two processes remains one of the very important research topics in the field of metaheuristic algorithms.In this study, a population-based metaheuristic algorithm named Tenrec Optimization Algorithm (TOA), inspired by the natural behaviour of the Tenrec animal in the wild, is introduced.Behaviour of this animal is investigated, and its characteristic strategies are extracted to design the proposed approach.The exploration phase is modeled with the inspiration of Random Foraging/Wide Area Searching behaviour and aims to increase the search space coverage, avoid repeating previous paths and guide agents towards unknown areas.Also, the exploitation phase is designed based on the digging behaviour in promising areas and adaptive return to the best past experiences to perform local search more accurately.The performance of TOA is evaluated using 23 standard benchmark functions including unimodal test functions, high-dimensional multimodal test functions and fixed-dimensional multimodal test functions.The results show that TOA has a high ability in exploitation in unimodal test functions, a good power in exploration in high-dimensional multimodal test functions and a good performance in balancing exploration and exploitation in fixed-dimensional multimodal test functions.Also, comparison with 9 well-known metaheuristic algorithms shows that TOA, by achieving the best rank in 20 functions out of a total of 23 benchmark functions, provides more competitive performance in about 86.95% of problems and can be used as an effective optimizer for a wide range of optimization problems.