Hands-On Genetic Algorithms with Python

Evaluating the solution

While evaluating a given solution, or when comparing two solutions, it is apparent that the reward value alone may not provide us with sufficient information. With the way the reward is defined, its value will always be -200 if we don't hit the flag. When we compare two candidate solutions that don't hit the flag, we would still like to know which one got closer to it and consider it a better solution. Therefore, we are going to use the final position of the car, in addition to the reward value, to determine the score of the solution. If the car did not hit the flag, the score will be the distance from the flag. Therefore, we will be looking for a solution that minimizes the score. If the car hits the flag, the base ...