Optimization-Based Agents in the 16th Multi-agent Programming Contest

Download PDF
Open Webpage

Paula Böhm

Published: 01 Jan 2022, Last Modified: 24 Jul 2025The Multi-Agent Programming Contest 2022EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: In this paper we present Blup, a system designed for the Agents Assemble III scenario of the 16th Multi-Agent Programming Contest (MAPC) that took place in 2022. In Blup, agents collaborate by choosing their actions together. This is implemented by two suitable constraint optimization problems, which are solved using Google OR-Tools. Blup won 11 out of 15 simulations in 5 games played outside the competition against all participants of the 16th MAPC, which is the second-best result achieved by any team in the contest when including the aforementioned games. We present our approach in detail, analyse the games played, and propose improvements.