The structure of adaptive competition in minority games

Manuca, Radu; Li, Yi; Riolo, Rick L.; Savit, Robert

doi:10.1016/s0378-4371(00)00100-x

Cited by 62 publications

(80 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1. The Correlation is highest for low m, consistent with the analysis in [25] showing herding behavior in this region. Within the low region, the mean value appears to increase from m = 1 to m = 2 before declining for m > 2, but given the size of the standard deviation in region, the most that can be said is that the Correlation is comparable for these two values.…”

Section: Correlation: Behavioral Joint Informationsupporting

confidence: 88%

“…As z decreases below this point, the system performance degenerates until the agents are doing worse than with random choices. In this region, the time series of minority groups [25] show "herding" behavior. Relatively few distinct strategies are available for small m, with high probability that agents' decisions will overlap.…”

Section: Correlation: Behavioral Joint Informationmentioning

confidence: 99%

See 1 more Smart Citation

A Design Taxonomy of Multi-agent Interactions

Parunak

Brueckner

Fleischer

et al. 2004

Agent-Oriented Software Engineering IV

View full text Add to dashboard Cite

Abstract. Agent interactions are frequently characterized as "coherent," "collaborative," "cooperative," "competitive," or "coordinated." These terms specialize the more foundational category of "correlation," which can be measured by the joint information of a system. "Congruence"is orthogonal to the others, reflecting the degree to which correlation and its specializations satisfy user requirements. A taxonomy of these mechanisms can guide the design of multi-agent interaction. Lack of correlation is sometimes necessary, and requires the use of formal stochasticity.

show abstract

Section: Correlation: Behavioral Joint Informationsupporting

confidence: 88%

Section: Correlation: Behavioral Joint Informationmentioning

confidence: 99%

A Design Taxonomy of Multi-agent Interactions

Parunak

Brueckner

Fleischer

et al. 2004

Agent-Oriented Software Engineering IV

View full text Add to dashboard Cite

show abstract

“…The periodic dynamics is a result of oversampling of strategy space and common zero initial conditions among agents when α < α c , accounting for the high volatility in the symmetric phase. It is reflected in a prominent period 2 M+1 peak in the autocorrelation of the attendance time series of a particular option [8,19,20,24,25]. Fig.…”

Section: Pure Population With Random Walk Approximationmentioning

confidence: 99%

“…Up to first order approximation, the dynamics of MG is determined by this control parameter α. [18,19,20] The most sparkling macroscopic observable in MG is perhaps the variance of option attendance per agents σ 2 /N . It represents the wastage of the system and fluc-tuation of resources allocation; the smaller σ 2 /N , the more the whole population benefits.…”

Section: Introductionmentioning

confidence: 99%

Incorporating inertia into multiagent systems

Man

Chau

2006

Phys. Rev. E

View full text Add to dashboard Cite

We consider a model that demonstrates the crucial role of inertia and stickiness in multi-agent systems, based on the Minority Game (MG). The inertia of an agent is introduced into the game model by allowing agents to apply hypothesis testing when choosing their best strategies, thereby reducing their reactivity towards changes in the environment. We find by extensive numerical simulations that our game shows a remarkable improvement of global cooperation throughout the whole phase space. In other words, the maladaptation behavior due to over-reaction of agents is removed. These agents are also shown to be advantageous over the standard ones, which are sometimes too sensitive to attain a fair success rate. We also calculate analytically the minimum amount of inertia needed to achieve the above improvement. Our calculation is consistent with the numerical simulation results. Finally, we review some related works in the field that show similar behaviors and compare them to our work.

show abstract

“…As z decreases below this point, the system performance quickly degenerates until the agents are doing worse than if they made random choices. In this region, study of the time series of minority groups [25] shows "herding" behavior. Relatively few distinct strategies are available for small m, with high probability that agents' decisions will overlap.…”

Section: Correlation: Behavioral Joint Informationmentioning

confidence: 98%

A preliminary taxonomy of multi-agent interactions

Parunak

Brueckner

Fleischer

et al. 2003

Proceedings of the Second International Joint Conference on Autonomous Agents and Multiagent Systems - AAMAS '03

View full text Add to dashboard Cite

Discussions of agent interactions frequently characterize behavior as "coherent," "collaborative," "cooperative," "competitive," or "coordinated." We propose a series of formal distinctions among these terms and several others. We argue that all of these are specializations of the more foundational category of "correlation," which can be measured by the joint information of a system. We also propose "congruence" as a category orthogonal to the others, reflecting the degree to which correlation and its specializations satisfy user requirements. Then we explore the degree to which lack of correlation can arise purposefully, and show the need to use formal stochasticity in cases where such lack of correlation is truly necessary (such as in stochastic search).

show abstract

The structure of adaptive competition in minority games

Cited by 62 publications

References 7 publications

A Design Taxonomy of Multi-agent Interactions

A Design Taxonomy of Multi-agent Interactions

Incorporating inertia into multiagent systems

A preliminary taxonomy of multi-agent interactions

Contact Info

Product

Resources

About