A Nature-Inspired Hybrid Heuristic for Orchestrating the Self-Deployment of Mobile Supply Robots in...

We present a nature-inspired approach for self-organizing logistics in contested environments characterized by uncertainty. The presented approach leverages territorial partitioning principles observed in natural collectives and systems. The logistics challenge addressed involves autonomously deploying a fleet of uncrewed mobile robots tasked with dynamically adjusting their positions to optimally service spatially and temporally fluctuating demands. Unlike traditional logistics, our approach is largely decentralized, relying almost exclusively on local interactions and decision-making. Monte Carlo simulations are conducted to evaluate performance across different scenarios, varying in client distribution (from uniform to highly clustered) and the range of the local perception of the logistic platform. Results demonstrate that the decentralized allocation strategy can deliver logistics support at a performance on par with a traditional clustering method, such as k-means, at runtime and without preliminary offline calculations.