Evaluation Function

• This is all dependent on the evaluation function.
• You need fitter agents to breed, so you need to say who is fitter.
• For the travelling salesman problem, the function is easy; you just calculate the length of the path (though you might want to do something clever with illegitimate agents).
• In general, implementing this evalution function is not straight forward.
• In nature, it's not really clear how to evaluate an agent.
• You just have to run the simulation (see who lives and reproduces).
• What you are really looking for is a good (or some good) solutions for the evaluation function.