Problem solving by heterogeneous agents (Q5937319): Difference between revisions

Consider a finite number of boundedly rational problem solvers (economic agents), each of whom is characterized by a perspective and a heuristic. This pair of characteristics induces a learning style, which the agent applies to problems faced. Suppose all agents face the same finite problem, and no two agents are alike in learning style. Then the agents, formed in binary strings, are sure to arrive at the maximum value of the problem. However, the path to the maximum value need not display expected regularities. For instance, with a concave problem, one may observe increasing returns to marginal problem solvers.