ABSTRACT
In the notional class of all agent systems, there are ones which have local interactions, and non-local interactions. When individual agents interact with one another, it's a social interaction, in the sense that's used by economists, or sociologists. Agents are often purposive. Typically they may have a utility function they want to maximize. An agent-based system can only be shown to be an optimal configuration not with respect to some kind of global optimality, but with respect to evaluating the welfare of individuals. The strength of agent-based modeling over equation-based modeling is that the model structure, being divided up according to entities in the world, as opposed to observables, is often advantageous in a commercial environment. Agents in general like to work, they like income, but they also like to engage in leisure time, time that they don't work.
