Improved remora optimization algorithm combining chaotic mapping and crossover strategy: path planning for orchard robots

Orchard robot path planning has become a major area of study in smart agriculture due to advances in intelligent robotics technology. To address limitations of the Remora Optimization Algorithm (ROA)—such as low convergence accuracy, susceptibility to local optima, and inadequate global search—this study introduces an enhanced algorithm (GROA) that employs Logistic-Tent chaotic mapping for population initialization, a crossover mechanism to improve search update, and a sine-cosine method to balance global and local exploration. Compared with ROA, Harris Hawks Optimization (HHO), and the Seahorse Optimization algorithm (SHO) using the CEC2020 benchmark, GROA achieves average improvements of 20% in ideal fitness, 32% in average fitness, and 73% in standard deviation. When applied to orchard path planning, GROA also outperforms ROA, Genetic Algorithm (GA), and Particle Swarm Optimization (PSO), with average gains of 0.23% in the shortest path and 3.26% in average path length.