Sample greedy based task allocation for multiple robot systems

Abstract This paper addresses in-schedule dependent task allocation problems for multi-robot systems. One of the main issues with those is inherent NP-hardness combinatorial optimisation. To handle this issue, develops a decentralised algorithm by leveraging submodularity concept and sampling process sets. Our theoretical analysis reveals that proposed can provide an approximation guarantee 1/2 optimal solution monotone submodular case 1/4 non-monotone case, both polynomial time complexity. examine performance validate analysis, we introduce two scenarios perform numerical simulations. The simulation results confirm achieves quality which comparable to state-of-the-art algorithms in much better significantly lower computational