Development of the Evolutionary Relay Rider Optimization Algorithm for a Robust Solution Over Time in Dynamic Optimization Problems

Based on the evolutionary Rider Optimization Algorithm (ROA), which has demonstrated its effectiveness for single-objective problems, a multi-population Relay Rider Optimization Algorithm (RROA) has been developed for dynamic optimization problems. The RROA incorporates several additional functions: local search for each rider’s optimal position, grouping and storing central rider positions, a relay mechanism among rider squads, an exclusion mechanism to prevent convergence to the same optimum, and maintaining squad diversity. To enhance robust solution survival time, the extended RROA calculates the next robust solution using peak shift severity, fitness function variance of the best solutions after environmental changes, and height variance of peak heights. The RROA’s performance was evaluated using a modified Moving Peak Benchmark and showed a 35% increase in average survival time for robust solutions compared to alternative algorithms like RFTmPSO-s4, RAmQSOs4, RmNAFSA-s4, Jin’s, Fu ( \(\theta = 0\) ), Fu ( \(\theta = 1\) ), Guo’s.