From Theory to Air Force Practice: Applications and Non-Binary Extensions of Probabilistic Model-Building Genetic Algorithms

摘要

These goals have been substantially accomplished. Moreover, new findings led to important discoveries that were unanticipated at the beginning of the project. The report starts with a retrospective examination of project accomplishment. 1) Develop, implement, and enhance probabilistic model-building GAs for non-binary codes. 2) Extend existing facet wise models to non-binary codes. 3) Extend bounding test functions to non-binary code. 4) Extend the proposed non-binary algorithms to hierarchically difficult problems. 5) Apply the developed algorithms to two problems of Air Force interest.