Example 2 - Extending the `El Farol Bar' Model

Description of the Original Model

Brian Arthur modelled this by dealing each agent a fixed number of models randomly allocated from a limited number of types, for example `the same as last week', `60 minus the number who went two weeks ago', `the average over the last 4 weeks', or `the number predicted by the trend over the last three weeks'. Then each week each agent evaluates all its models against the past record of how many went and find the best predictive model. It then uses this model to predict the number who will go this week and bases its decision on this. All the agents do this simultaneously and in parallel.

The resulting number who go each week seems to oscillate stochastically about the critical 60% mark, similar to the shape in that in figure 1, despite the fact that this model once initialised is strictly deterministic. Each agent's repetoire of models is fixed and only the current assessment of the models changes from week to week, so that at different times different models will be most successful. The only communication which takes place is the implicit message of the number who go each week.

Figure 1. Number of people going to El Farol's each week in a typical run (with 10 agents)

One of the interesting things in this model is that although each agent is dealt a different menu of models, and all decide to go or not in different combinations and at different times, at the level of whole model they are pretty indistinguishable in terms of their behaviour.

Generated with CERN WebMaker