An important capacity of an agent in a multi-agent world is the ability to define one<<s own goals. The agent<<s goals are often affected by (and in some cases determined by) the goals of other agents.
The process of goal definition seems to subsume goal selection. Mitchell (1990) has defined perception as a process linking the state of the world to the appropriate goals to attend to. This process involves selecting the most pertinent goal and trying to achieve it in preference to others.
So far not much work has been done in the area of goal definition. Most work done has concentrated on goals of two agents only. Baker<<s system KANT (1991), for example, incorporates reasoning mechanisms for determining which set of goals are to be negotiated in a tutorial interaction.
When considering the relationships between individual goals and community goals two issues arise. First, how the satisfaction of individual goals affects the satisfaction of community goals. Then, how the satisfaction of community goals leads to the satisfaction of individual goals.
It is also possible to envisage that agents could learn which goals to pursue in order to achieve some overall goals. Perhaps the agents would follow a scheme of gradual differentiation that is common in human society. The agents start with similar goals, but differences in local conditions and agent-specific skills gradually differentiate the agents<< goals so as to function better in a community. Ideally this process lets the community evolve from a fairly uniform group to a differentiated highly competent society.