Minimising inequality in multiagent resource allocation

Abstract We analyse the problem of finding an allocation resources in a multiagent system that is as fair possible terms minimising inequality between utility levels enjoyed by individual agents. use well-known Atkinson index to measure and we focus on distributed approach resource allocation, where new allocations emerge result sequence local deals groups agents who agree exchange some items their possession. Our results show it design systems provide theoretical guarantees for optimal outcomes minimise inequality, but also there are significant computational hurdles be overcome worst case. In particular, computationally intractable under large number structurally complex deals, possibly involving many items, may required before convergence socially allocation. This remains true even severely restricted scenarios all have same function. From methodological point view, while much work relies combinatorial arguments, here instead insights from basic calculus.