Abstract In Multi-Agent (combined (MAPF), a team Findingtarget-assignment needs finding) to findfor problem collision-free teams of agents paths infromknow t cationsboth generalizes therespective to their anonymous and Combined targets. non-anoagent signment path-finding and Path Finding Each problems. (TAPF)ofextends th including the same number the of targetsofasassigning problem there are targe ag precursor Each aagent has toto the move MAPF problem.one to exactly A lim tteam models such that all assumption is their targets arethat visited. The the number targets is to first are equal, assign agents which is invalidand to targets the in somefree paths for thethis We address limitation agents by generalizing to their targets in T for (1) unequal number ofWe agents tas and the the makespan is minimized. present have deadlines by which they must Based Min-Cost-Flow) algorithm, a hierarchical algorithm be com dering TAPF that solves of groups of tasks to instances be completed; optimally by that are composed of a sequence of checkpoints that mustfrom anonymous and non-anonymous multi-agent pathbe visited in a specific order. Further, we model the pr using answerOn findinglemalgorithms. set the low level,(ASP) programming C cost max-flow algorithm customizing the desiredon a time-expanded variant of the prob assign– all agents in a single team to one only needs to choose the appropriate combination their of paths. ASP rulesOn to theenforce high it. level, CBM We also us demon searchmentally to resolve thatcollisions if problemamong specificagents i informati corporatedwe Theoretically, into the ASP prove thatencoding CBM isthen AS correc ods can optimal. be efficient andwecan Experimentally, scalethe show to up sca to TAPFapplications. instances with dozens of teams and hundreds of agents and adapt it to a simulated warehouse system.