Chaotic dual ABC and chaotic Bee Wolf Optimizer for power system problems: A case study on electric vehicle charging station allocation

This paper proposes two advanced metaheuristic frameworks, the Chaotic Dual Artificial Bee Colony (CDABC) and the Chaotic Bee Wolf Optimizer (CBWO), to address the limitations of the standard Artificial Bee Colony (ABC) algorithm. While standard ABC is effective for global optimization, it often suffers from slow convergence and a tendency to become trapped in local optima. The proposed CDABC integrates a chaotic sine map to generate a time-varying, adaptive step-size factor, combined with an Antithetic Dual Search (ADS) strategy that generates two neighbour candidates in opposite directions to enhance local refinement. To further enhance global exploration, the CBWO hybridizes the CDABC framework by replacing the employed bee phase with the hunting mechanism of the Grey Wolf Optimizer (GWO), utilizing alpha, beta, and delta wolf operators to guide the population. This optimization is then divided into two distinct phases. First, algorithmic efficiency and numerical stability are validated using five diverse benchmark functions. Second, the algorithms are applied to a practical power system problem, which is the optimal allocation of Electric Vehicle Charging Stations (EVCS) within the IEEE 69-bus distribution network. The application minimizes total power loss and Total Voltage Deviation (TVD) as independent objectives. Results indicate the proposed algorithms outperform standard ABC and GWO in accuracy and consistency. Over 30 independent trials, all frameworks successfully converged near the optimal power loss boundary; however, CDABC demonstrated superior computational stability. It achieved a tighter standard deviation in active power loss (6.49e-03) compared to the standard ABC (8.20e-03) and minimized Total Voltage Deviation with higher precision, yielding a lower standard deviation across runs (3.11e-07). These specific statistical metrics demonstrate that the chaotic dual search mechanism successfully mitigates stochastic volatility, establishing a highly robust optimization environment. Ultimately, these findings confirm that the proposed hybrid frameworks provide a highly effective solution for network operators, ensuring that future EVCS expansion can be accommodated while actively minimizing energy dissipation and preserving the TVD within safe operational limits.