<scp>MIFTTA</scp>: Multi‐Strategy Improved Football Team Training Optimization Algorithm for Engineering Applications

ABSTRACT The football team training algorithm (FTTA) is a novel meta‐heuristic optimization technique inspired by the three stages of a football team's training process: collective, group, and individual training. Although FTTA exhibits good competitiveness in comparison with other algorithms, it still has a number of drawbacks, including slow convergence speed, low convergence accuracy, insufficient perturbation, and a propensity to enter local optima when solving some complex problems with high dimensional and non‐linear constraints. To address these drawbacks, this paper introduces an improved variant of FTTA, termed multi‐strategy improved football team training algorithm (MIFTTA). First, an adaptive bilateral factor is introduced to effectively balance the global exploration and local exploitation capabilities of the algorithm. Second, an adaptive oscillating inertia weighting factor is implemented to accelerate the convergence process. Then, building on the adaptive cluster grouping mechanism of the original algorithm, an inter‐group communication mechanism is integrated to enhance population diversity during the convergence process, thereby improving the convergence accuracy. Finally, a population bi‐directional restart mechanism is devised to strengthen the algorithm's ability to escape from the local optima and explore the solution space more comprehensively. To validate the overall performance of MIFTTA, it is compared with various state‐of‐the‐art algorithms in the CEC2017 and CEC2022 benchmark suites. The results show that MIFTTA achieves average rankings of 1.48 and 2.08 on the two test suites, respectively, with an overall final rank of 1. In the majority of test cases, MIFTTA provides more accurate and reliable solutions than other competitors. Furthermore, MIFTTA is applied to six real‐world engineering optimization problems and two photovoltaic model parameter identification problems. The experimental results demonstrate that MIFTTA outperforms the competing algorithms in terms of solution quality and computational efficiency, showing its potential for solving complex optimization problems.