Teaching-learning-based pathfinder algorithm for function and engineering optimization problems

Pathfinder algorithm (PFA) for finding the best food area or prey based on the leadership of collective action in animal groups is a new metaheuristic algorithm for solving optimization problems with different structures. PFA is divided into two stages to search: pathfinder stage and follower stage. They represent the exploration phase and mining phase of PFA respectively. However, the original algorithm also has the problem of falling into a local optimum. In order to solve this problem, the teaching phase in the teaching and learning algorithm is added to the pathfinder stage in the text. In order to balance the exploration and mining capabilities of the algorithm, the learning phase of the teaching and learning algorithm is added to the follower phase in the article. In order to further enhance the depth search ability of the algorithm and increase the convergence speed, the exponential step is given to the followers. Therefore, a teaching-learning-based pathfinder algorithm (TLPFA) is proposed. 19 benchmark functions of four different types and six engineering design problems are used to test of the TLPFA exploration and exploiting capabilities. The experimental results show that the proposed TLPFA algorithm is superior to the state-of-the-art metaheuristic algorithms in terms of the performance measures.