Efficient and reliable scheduling of power generating units in the unit commitment problem using the Tardigrade optimization algorithm

Abstract

Introduction. The unit commitment (UC) problem is a critical operational task in power systems, involving the optimal scheduling of generating units while meeting demand, satisfying technical constraints, and minimizing operating costs. Due to its combinatorial nature, nonlinear characteristics, and numerous interdependent constraints, UC poses a highly complex optimization challenge. Metaheuristic algorithms have demonstrated strong potential in addressing such large-scale problems; however, many existing methods struggle to maintain a proper exploration–exploitation balance, limiting their performance in dynamic UC environments. Problem. Traditional metaheuristic algorithms often suffer from premature convergence, inadequate local refinement, or dependency on control parameters that require tuning. Such limitations reduce robustness and adaptability when dealing with UC’s intricate search landscape. Therefore, there is a need for a parameter-free, self-adaptive optimization algorithm capable of reliably solving UC with high efficiency and convergence stability. The goal of this study is to develop an efficient and reliable scheduling framework for power generating units in the UC problem by employing the tardigrade optimization algorithm (TOA) and to demonstrate its effectiveness compared with established optimization techniques. Methodology. TOA is inspired by the active and cryptobiotic survival behaviors of tardigrades. The exploration phase imitates active adaptive locomotion to broaden global search, while the exploitation phase abstracts cryptobiotic stability to refine solutions locally. These mechanisms are formulated through adaptive state-transition operators that adjust search behavior automatically without external parameters. TOA is applied to a 24-hour UC problem consisting of 10 generating units under realistic load and operational constraints. Its performance is benchmarked against 6 widely used metaheuristic algorithms. Results. The proposed TOA achieves the lowest total operating cost, exhibits strong convergence behavior, and demonstrates high consistency across independent runs, outperforming all comparative methods. The scientific novelty lies in introducing a biologically inspired, parameter-free, self-adaptive metaheuristic algorithm. Its practical value is validated through superior performance in UC scheduling, indicating strong potential for broader power system optimization tasks. References 21, tables 3, figures 3.