Social Capital as a Complexity Reduction Mechanism for Decision Making in Large Scale Open Systems

摘要

A common requirement of distributed multi-agent systems is for the agents themselves to negotiate pairwise agreements on performing a joint action. In systems with endogenous resources, the cost of computing the decision-making has to be taken into account. If the computational resources expended in negotiating an optimal solution exceed the marginal benefits gained from that negotiation, then it would be more expedient and efficient to use the memory of past interactions to short-cut the complexity of decision-making in joint or collective actions of this kind. In social systems, it has been observed that social capital is an attribute of individuals that enhances their ability to solve collective action problems. In this paper, we define a new computational framework for representing and reasoning about electronic social capital, in which actions enhance or diminish three different forms of social capital (individual trustworthiness, social network, and institutions), and a decision-making model which uses social capital to decide whether to cooperate or defect in strategic games. A set of scenarios are presented where we believe that social capital can support effective collective action sustained over time, avoid suboptimal dominant strategies, and short-cut the computational costs involved in repetitious solving of strategic games.