CEC Tutorial'07

Historical roots:Evolutionary Computation:A Unified Approach • Evolution Strategies (ESs):– developed by Rechenberg, Schwefel, etc. in 1960s.Kenneth De Jong– focus: real-valued parameter optimization– individual: vector of real-valued parametersComputer Science DepartmentGeorge Mason University – reproduction: Gaussian “mutation” of parameterskdejong@gmu.edu– M parents, K>>M offspringwww.cs.gmu.edu/~eclab1 2Historical roots: Historical roots:• Genetic Algorithms (GAs):• Evolutionary Programming (EP):– developed by Holland in 1960s– Developed by Fogel in 1960s– goal: robust, adaptive systems– Goal: evolve intelligent behavior– used an internal “genetic” encoding of points– Individuals: finite state machines– reproduction via mutation and recombination of– Offspring via mutation of FSMsthe genetic code.– M parents, M offspring– M parents, M offspring3 4Present Status: Interesting dilemma:• wide variety of evolutionary algorithms (EAs)• A bewildering variety of algorithms and• wide variety of applicationsapproaches:– optimization– GAs, ESs, EP, GP, Genitor, CHC, messy– search GAs, …– learning, adaptation• well-developed analysis • Hard to see relationships, assess strengths& weaknesses, make choices, ...– theoretical– experimental5 6Viewpoint:A Personal Interest:• Develop a general framework that:?– Helps one compare and contrast approaches.– Encourages crossbreeding.GA ES EP GP . . .– Facilitates intelligent design choices.7 ...