4 Towards dealing with the complexity of modelling agents - modelling modelling

4.1 The Form - meaning distinction

The most fundamental distinction, as discussed above, is between the form of an agent's model and its meaning. Where the meaning of a model is determined by the mapping of the input and resulting predictions to the model. This is very similar to the syntax-semantics distinction in formal logic*1. This distinction is critical where we are dealing with agents of bounded rationality. For example, although two models may be equivalent in terms of its predictions, one could be so costly to use that it may be all but useless in deciding on an appropriate action.

In a very general way this can be formalised by reference to the space of all relevant possibilities (similar to phase space diagrams used in physics). Models, goals, a priori knowledge, observations, goals and actions can all be associated with subspaces of this (see figure 2).

Figure 2: Form and meaning distinction

From Complexity to Agent Modelling and Back Again - Bruce Edmonds - 15 MAY 97

[Next] [Previous] [Up] [Top] [Contents]

Generated with CERN WebMaker