Competition of Distributed and Multiagent Planners (CoDMAP)

In multi-agent planning, preserving the agents’ privacy has become an increasingly popular research topic. For preserving the agents’ privacy, agents jointly compute a plan that achieves mutual goals by keeping certain information private to the individual agents. Unfortunately, this can severely restrict the accuracy of the heuristic functions used while searching for solutions. It has been recently shown that, for centralized planning, the performance of goal oriented search can be improved by combining goal oriented search and width-based search. The combination of these techniques has been called best-first width search. In this paper, we investigate the usage of best-first width search in the context of (decentralised) multi-agent privacy-preserving planning, addressing the challenges related to the agents’ privacy and performance. In particular, we show that best-first width search is a very effective approach over several benchmark domains, even when the search is driven by heuristics that roughly estimate the distance from goal states, computed without using the private information of other agents. An experimental study analyses the effectiveness of our techniques and compares them with the state-of-the-art.

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Best-First Width Search for Multi Agent Privacy-Preserving Planning

Gerevini

Lipovetzky

Percassi

et al. 2019

“…We implemented our compilation, based upon the script which was used to convert MA-PDDL to PDDL representing a centralized planning problem from the first Competition of Distributed and Multiagent Planners (Stolba, Komenda, and Kovacs 2015). 2 We remark that while we have described the compilation for a grounded MA-STRIPS setting, most of the compilation can actually be done on the lifted level of MA-PDDL, and our compilation performs partial grounding where it must.…”

Section: Empirical Evaluationmentioning

confidence: 99%

Automated Verification of Social Law Robustness in STRIPS

Karpas

Shleyfman

Tennenholtz

2017

Agents operating in a multi-agent environment must consider not just their own actions, but also those of the other agents in the system. Artificial social systems are a well known means for coordinating a set of agents, without requiring centralized planning or online negotiation between agents. Artificial social systems enact a social law which restricts the agents from performing some actions under some circumstances. A good social law prevents the agents from interfering with each other, but does not prevent them from achieving their goals. However, designing good social laws, or even checking whether a proposed social law is good, are hard questions. In this paper, we take a first step towards automating these processes, by formulating criteria for good social laws in a multi-agent planning framework. We then describe an automated technique for verifying if a proposed social law meets these criteria, based on a compilation to classical planning.

“…For the experimental evaluation, we use the distributed track setup of the CoDMAP 5 ( Štolba, Komenda, and Kovacs 2016) competition including all 12 benchmarks. Each agent runs on separate machine with i5-4460 3.4GHz processor and 8GB memory and has its own problem and domain input files.…”

Section: Experimental Evaluationmentioning

confidence: 99%

Potential Heuristics for Multi-Agent Planning

Štolba

Fišer

Komenda

2016

Self Cite

Distributed heuristic search is a well established technique for multi-agent planning. It has been shown that distributed heuristics may crucially improve the search guidance, but are costly in terms of communication and computation time. One solution is to compute a heuristic additively, in the sense that each agent can compute its part of the heuristic independently and obtain a complete heuristic estimate by summing up the individual parts. In this paper, we show that the recently published potential heuristic is a good candidate for such heuristic, moreover admissible. We also demonstrate how the multi-agent distributed A* search can be modified in order to benefit from such additive heuristic. The modified search equipped with a distributed potential heuristic outperforms the state of the art.