Multi-objective Robot Coalition Formation for Non-additive Environments

Research towards the coalition formation problem in multi-robot systems has recently gained attention. The main objective of this research is to form the best teams of heterogeneous robots (coalitions) to cater to a given set of tasks. Due to the inherently NP-hard nature of the problem, it is being addressed employing a large number of techniques ranging from heuristics based solutions to evolutionary approaches. The problem becomes more complex when the resource-distribution needed to complete a task is non-additive in nature, i.e., it may not be sufficient to just add the resources to the individual robots forming the coalition to sum up to the resource requirement of the given task but also satisfy the minimum resource distribution on each individual member of the coalition. There may be multiple alternate coalitions for a task, each of which can complete the task but with different efficiency. Traditionally the coalition formation problem has been formulated as a single objective optimization problem wherein the objective is either to maximize the number of the tasks or to maximize the overall system efficiency. In this paper, the coalition formation problem has been modeled as a multi-objective optimization problem where both the number of tasks completed as well as the overall system efficiency are considered simultaneously. Two popular multi-objective approaches are implemented and the results demonstrate their superiority over single objective solutions.