Evolution of Cooperative Problem Solving in an Artificial Economy

Baum, Eric B.; Durdanovic, Igor

doi:10.1162/089976600300014700

Cited by 15 publications

(10 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…That is, the learning of individual self-interested agents and the learning of cooperation among these self-interested agents are simultaneous and thus interacting. This model extends existing work, in that it is not limited to bidding alone, for example, not just bidding alone for forming coalitions (as in Rosenschein and Zlotkin 1994) or bidding alone as the sole means for learning (as in Baum and Durdanovic 2000). Neither is it a model of pure reinforcement learning, without explicit interaction among agents (such as Shoham and Tennenholtz 1994, Hu and Wellman 1998, Littman 2001.…”

Section: Introductionmentioning

confidence: 76%

“…This value is given to the Q module so that it can take this payoff into account when deciding on its course of action (e.g., whether to reach one giving-up point or another). 4 Therefore, summarizing the above two learning rules, a Q value of an agent (for a particular state-action pair) is the expected (discounted) total reinforcement that the agent will receive from that point on. The Q module of an agent then decides the actions of the agent based on maximizing the expected (discounted) total reinforcement that the agent will receive.…”

mentioning

confidence: 99%

See 1 more Smart Citation

MARLBS: Team Cooperation through Bidding

Sun¹,

Qi²

2005

IJCIR

View full text Add to dashboard Cite

Abstract:A cooperative team of agents may perform many tasks better than isolated agents. The question is how cooperation among self-interested agents may be achieved. It is important that, while we encourage cooperation among agents to form a team, we maintain autonomy of individual agents as much as possible, so as to maintain flexibility and generality. This paper presents an approach toward this goal, based on bidding utilizing reinforcement values acquired through reinforcement learning. The result is a simple and straightforward method that is generic and works in a variety of task domains. We further apply evolutionary computation to enhance cooperation among agents of a team, through selecting and reproducing those teams that are able to cooperate. We tested and analyzed this approach, MARLBS, in a variety of task domains, and demonstrated that a team of self-interested agents indeed performed better than the best single agent as well as the average of the single agents. In particular, Backgammon players trained using this approach outperformed PubEval (a publicly available benchmark player). These results validated our approach.

show abstract

Section: Introductionmentioning

confidence: 76%

mentioning

confidence: 99%

MARLBS: Team Cooperation through Bidding

Sun¹,

Qi²

2005

IJCIR

View full text Add to dashboard Cite

show abstract

“…Thus in his model each model has an implicit domain, in that it is only applies when it out-bids other models in order to be applied (Baum and Durdanovic, 2000b). In the most recent version of his algorithm (called Hayek 4) he also introduces explicit conditions of application as each model is a Post production rule (Baum and Durdanovic, 2000a).…”

Section: Evolutionary Computationmentioning

confidence: 99%

Learning Appropriate Contexts

Edmonds

2001

SSRN Journal

View full text Add to dashboard Cite

Abstract. Genetic Programming is extended so that the solutions being evolved do so in the context of local domains within the total problem domain. This produces a situation where different "species" of solution develop to exploit different "niches" of the problem -indicating exploitable solutions. It is argued that for context to be fully learnable a further step of abstraction is necessary. Such contexts abstracted from clusters of solution/model domains make sense of the problem of how to identify when it is the content of a model is wrong and when it is the context. Some principles of learning to identify useful contexts are proposed.

show abstract

“…An example of this is the Artificial Economy model [7] based on an economic paradigm, which acts as an evolutionary rule-based system for sequential decision tasks. Comparisons are made between the economic system and early classifiers, however it was developed separately from the improved classifier systems ZCS and XCS.…”

Section: Introductionmentioning

confidence: 99%

An activation reinforcement based classifier system for balancing generalisation and specialisation (ARCS)

Knittel

2010

Proceedings of the 12th Annual Conference Companion on Genetic and Evolutionary Computation

View full text Add to dashboard Cite

Learning Classifier Systems are reinforcement-based learning systems that allow the development of generalised rule sets. They allow a balance between higher level generalisable learning and reinforcement, and have been used in a number of systems to introduce principles from psychology to guide methods of learning. A classifier system based on Activation Reinforcement (ARCS) is described, based on accessibility of traces in semantic memory, that provides a strength related learning technique allowing balance of generalisation and specialisation of rules. The system is based on a minimal number of design principles, is arguably simpler in design than existing classifier systems, and has clearer connections with human cognitive models. Performance on the standard Woods environments shows fast, stable learning allowing near-optimal behaviour. The methods used by ARCS have a number of differences with existing ZCS and XCS approaches, that have potential advantages regarding generalisation techniques, and for addressing sparsely rewarded domains.

show abstract

Evolution of Cooperative Problem Solving in an Artificial Economy

Cited by 15 publications

References 10 publications

MARLBS: Team Cooperation through Bidding

MARLBS: Team Cooperation through Bidding

Learning Appropriate Contexts

An activation reinforcement based classifier system for balancing generalisation and specialisation (ARCS)

Contact Info

Product

Resources

About