Convergence to correct solutions in Genetic Algorithms depends largely on the fitness function. A fitness function that captures all goals and constraints can be difficult to find. This paper gives a mathematical justification for a fitness function that has previously been demonstrated experimentally to be effective.