Abstract
Coot algorithm (COOT) is a new metaheuristic algorithm (MA) based on the swarm of COOT birds' behavior. Like other MAs, the COOT algorithm may easily become trapped in local optima, tends to low diversity, slow convergence speed, and imperfect balance between exploitation/exploration. Thus, to overcome these shortcomings of the original COOT, seven effective strategies are injected into COOT, such as - best agent guide, control randomization, Transition Factor (TF), adjusting the position, leading the group towards the optimal area, phasor operator and Opposition Based Learning (OBL). To be specific, to alleviate these shortcomings, a new modified Coot algorithm (COOT) called mCOOT algorithm is proposed. To investigate the efficiency of mCOOT, it has been verified on complex and standard test suite given in IEEE CEC'2017 and its performance are compared with seven stable well-known MAs. In addition, the Wilcoxon sign rank test has been used to check the results' significance. To ensure the robustness of mCOOT, it has been used to identify the optimal lithium-ion (Liion) battery model parameter. Battery parameters are unmeasured parameters that can only be estimated. The main advantage of this identification algorithm is the high accuracy where the root mean square error (RMSE) has been reduced to 7.11548 x 10-4 with an efficiency of 99.25% and STD of 0.59466 x 10-5. Thus, we observed that mCOOT algorithm can be used as an efficient algorithm to solve electrical and complicated optimization problems.