Answers are grounded in this paper and cite their sources.
Nature-inspired metaheuristic algorithms are widely used for solving complex continuous optimization problems; however, maintaining a reliable balance between broad exploration and stable local refinement remains challenging, particularly under fixed computational budgets and high-dimensional search conditions. This paper proposes a new physics-inspired population-based optimizer, called Shadow-Based Optimization (ShBO), motivated by the geometric relationship between solar altitude and shadow length. In ShBO, a simplified solar-geometry model maps the optimization process onto a synthetic daylight interval and computes a shadow-length term from the solar altitude angle. This term serves as an adaptive controller of the movement magnitude. The algorithm employs two complementary search regimes: a stochastic exploration phase in which Gaussian random directions are scaled by the shadow-length mechanism, and an exploitation phase in which the movement direction is determined by the best-so-far solution, a randomly selected peer solution, and a controlled Gaussian perturbation. The shadow-length term and the iteration-dependent shrinkage component continuously modulate the movement magnitude throughout the search, while the algorithm switches from the exploration to the exploitation regime at the predefined phase boundary. The performance of ShBO is evaluated on the CEC2017 benchmark suite with dimensions D = 10 , 30, 50, and 100, and on the CEC2020 benchmark suite with D = 10 and 20, using 30 independent runs under a fixed budget of 30,000 objective-function evaluations per run. To assess practical applicability, ShBO is further tested on five constrained engineering design problems: the pressure vessel, tension–compression spring, three-bar truss, cantilever beam, and tubular column design problems, under the same 30,000-NFE budget. In addition, ShBO is applied to photovoltaic parameter estimation using RTC France I–V measurements at 33 °C for the single-diode, double-diode, and triple-diode models, using a fixed budget of 50,000 NFEs per run. In the photovoltaic application, the implicit current equation is solved by the Newton–Raphson method at each voltage sample, and the objective function is defined as the RMSE between measured and estimated currents. The results are analyzed using mean, standard deviation, best, worst, Friedman mean rank, convergence behavior, boxplots, and Wilcoxon signed-rank tests at α = 0.05 . The experimental findings indicate that ShBO achieves competitive performance across the evaluated numerical benchmark functions and constrained engineering problems relative to the comparison algorithms considered in this study. In the photovoltaic application, the estimated models yield low RMSE values with limited run-to-run variability and close agreement with the measured I–V and P–V characteristics. These results indicate the applicability of the proposed shadow-length-based search mechanism to numerical optimization, constrained engineering design, and nonlinear parameter-estimation tasks under the experimental settings considered in this study.
No takes yet. Share an insight, caveat, or question.
Altay et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: