Cooperative Coevolutionary Genetic Algorithm for Multirobot Task Scheduling in Antarctica region☆

초록

Multi-Robot Task Scheduling (MRTS) is a critical challenge in deploying Multi-Robot Systems (MRSs) for complex missions such as exploration, search and rescue, and surveillance. MRTS is an NP-hard problem, necessitating efficient approximation algorithms. Traditional Genetic Algorithms (GAs) often struggle with scalability in large MRTS instances due to limitations in representation and chromosome design. Single-chromosome encodings perform well for small problem sizes but scale poorly, while na & iuml;ve multi-chromosome variants fare no better. Furthermore, Antarctica's extreme climate and icy terrain severely reduce robotic structural integrity and mechanical reliability, while also hampering mobility due to reduced traction. To ensure reliable operation, task scheduling algorithms must incorporate terrain factors such as elevation gradients and slopes into their cost functions, alongside distance metrics. To address this, we embed the multi-chromosome encoding into the Cooperative Coevolutionary Genetic Algorithm (CCGA) framework to jointly optimize task allocation and scheduling. Specifically, we introduce an iterative vector (iVec) clustering mechanism that partitions tasks based on radial distance and angular bearing from a common depot, ensuring geometrically coherent subproblems and balanced workloads. Each subproblem is then optimized in parallel by an independent GA subpopulation. Furthermore, we incorporate a dynamic decomposition update mechanism using Delaunay triangulation for efficient task reallocation, enhancing adaptability to dynamic environments and preventing convergence to local optima. Experimental results across synthetic and real-world Antarctic environments demonstrate that our proposed iVec-CCGA consistently outperforms existing benchmark algorithms, including single-chromosome GAs, multi-chromosome GAs, and other cooperative coevolutionary approaches, particularly in larger-scale and more complex scenarios.

키워드