Slime Mould Inspired Models for Path Planning: Collective and Structural Approaches

Abstract

Path planning is a classic and important problem in computer science, with manifold applications in transport optimisation, delivery scheduling, interactive visualisation and robotic trajectory planning. The task has been the subject of classical, heuristic and bio-inspired solutions to the problem. Path planning can be performed in both non-living and living systems. Amongst living organisms which perform path planning, the giant amoeboid single-celled organism slime mould Physarum polycephalum has been shown to possess this ability. The field of slime mould computing has been created in recent decades to exploit the behaviour of this remarkable organism in both classical algorithms and unconventional computing schemes. In this chapter we give an overview of two recent approaches to slime mould inspired computing. The first utilises emergent behaviour in a multi-agent population, behaving in both non-coupled and coupled modes which correspond to slime mould foraging and adaptation respectively. The second method is the structural approach which employs numerical solutions to volumetric topological optimisation. Although both methods exploit physical processes, they are generated and governed using very different techniques. Despite these differences we find that both approaches successfully exhibit path planning functionality. We demonstrate novel properties found in each approach which suggest that these methods are complementary and may be applicable to application domains which require structural and mechanical adaptation to changing environments.