Local Optimization


See the discussions in the Terminology section for an explanation of the purpose of local optimization.


Description of the properties of the local optimization section of the Genetic Algorithm Tab in the Optimization Controller:



Fitness Evaluations


The number of ‘tries’ that the local optimizer will make before completing the generation. The higher the number of fitness evaluations, the more effective the local optimizer will be. However, the run time will also be higher.


The default value is 30.





The local optimizer method for the hill climbing. Available options are ‘Steepest Ascent’, ‘Next Ascent’, ‘Random Mutation’ and ‘Adaptive’. Refer to online sources for explanations of these algorithms.


The default method is Random Mutation.



Waiting Time


The number of generations that must pass without improvement of the best coverage before the local optimizer is used.


The default is set to 5.