摘要: The dynamical principle for a population of interacting individuals with mutual pairwise knowledge, presented by the author in a previous paper for the case of constant knowledge, is extended to include the possibility that the kn... 展开 The dynamical principle for a population of interacting individuals with mutual pairwise knowledge, presented by the author in a previous paper for the case of constant knowledge, is extended to include the possibility that the knowledge is time-dependent. Several mechanisms are presented by which the mutual knowledge, represented by a matrix K, can be altered, leading to dynamical equations for K(t). The author presents various examples of the transient and long time asymptotic behavior of K(t) for populations of relatively isolated individuals interacting infrequently in local binary collisions. Among the effects observed in the numerical experiments are knowledge diffusion, learning transients, and fluctuating equilibria. This approach will be most appropriate to small populations of complex individuals such as simple animals, robots, computer networks, agent-mediated traffic, simple ecosystems, and games. Evidence of metastable states and intermittent switching leads them to envision a spectroscopy associated with such transitions that is independent of the specific physical individuals and the population. Such spectra may serve as good lumped descriptors of the collective emergent behavior of large classes of populations in which mutual knowledge is an important part of the dynamics. 收起