When All Agents Die: Analyzing the “Failures” in an Agent-Based Model of Human Foraging

Abstract

When running a simulated social-historic scenario, we often find situations in which all agents die, even though the simulated population appears to grow in the first steps. Is this a signal that something is wrong in the computer model or its implementation? We analyze this issue in our computer model of cooperation and cultural diversity among hunter-gatherers in prehistory. We have calculated more than 11,000 possible parameter combinations, taking into account the growth and decay of the population and the availability of resources in the environment. When the initial population is too scarce or too big for the local availability of resources, it begins to decrease until it disappears. This can be a very trivial test for the Malthus condition, but we have discovered that there are other important correlations affecting social and economic factors that should be explored.