Multi-robot Task Allocation Problem: Current Trends and New Ideas

The use of robotic fleets, a.k.a. multi-robot computing systems, is a de–facto standard in a variety of real-world applications. Efficiently solving the multi-robot task allocation (mrta) problem is perhaps one of the most fundamental feature of such systems. In this paper, we attempt at overviewing recent research on the matter. In particular, we first briefly survey some of the most interesting results of the last decade. Then, we focus on a prominent variant, namely the singletask single-robot task allocation problem with hard temporal constraints (st-sr-hc, in short), and on auction-based algorithms, which have been shown to be rather effective solutions in the st-sr-hc setting. To this regard, we discuss the state-of-the-art and highlight some ideas we are currently exploring to improve the best algorithm of the category, i.e. tessi. We provide preliminary experimental evidences of how such ideas seem to lead to better balancing number of performed tasks, distance and makespan optimization objectives.