Games meet Concurrency: Algorithms and Hardness

Abstract

Since the turn of the 21st century, seeing the decline of Moore’s Law on the horizon, the pursuit of continued software performance gains has led to the prominence of computer architectures with high degrees of parallelism and memory cache hierarchies. However, there are still many challenges to designing efficient algorithms and understanding the complexity of fundamental problems in these new models of computation. Given the similarities of concurrent systems of multiple agents and multiplayer games, this thesis analyzes a spectrum of models connecting these three fields and bridges the gaps between them by building upon techniques from the growing literature studying the complexity of games through gadget motion planning frameworks.