Using Answer Set Programming for Assigning Tasks to Computing Nodes

Allocating tasks to computing nodes in a network is an important configuration problem. In the case of fail-safe networks, such configuration must be changed during operation if a computing node fails. Hence, a fast configuration is required. In this paper, we formulate a tasks-to-computing-nodes assignment problem and its constraints using answer set programming. We performed an initial experimental evaluation utilizing several smaller to mid-size problem instances to show whether logic reasoning based on answer set programming is feasible for practical applications. We discovered that reasoning is fast if a solution exists but not when there is no solution. Further constraints help to decide that a problem instance is unsolvable early in the search, which improves the outcome.