Computational benchmarking demonstrates superior global optimization and engineering performance for an enhanced Escape algorithm, highlighting its efficacy in complex search spaces.
Key Points
To overcome the limitations of slow convergence and premature convergence to local optima in the crowd-evacuation-inspired Escape algorithm.
Developed ESCSA by integrating a dynamic mask probability adjustment mechanism, stagnation detection with Simulated Annealing local search (Gaussian perturbation and Metropolis criterion), and a worst-elimination population replacement strategy.
Evaluated performance against 11 comparative algorithms across 10, 30, 50, and 100 dimensions of the CEC 2017 benchmark suite using the Friedman mean rank test.
Assessed real-world optimization capabilities on three constrained engineering design problems and mobile robot path planning in static environments.
Achieved first place in the Friedman mean rank test across all dimensions (mean ranks of 2.217, 1.383, 1.400, and 1.933 for 10, 30, 50, and 100 dimensions, respectively), securing optimal mean fitness on 17, 22, 21, and 17 functions.
Ranked second in speed reducer design, third in multi-disc clutch braking, and first in rolling element bearing design optimization.
Successfully navigated static environments by efficiently planning optimal paths for mobile robots.