Conference Paper

The Effect of Building Block Construction on the Behavior of the GA in Dynamic Environments: A Case Study Using the Shaky Ladder Hyperplane-Defined Functions.

DOI: 10.1007/11732242_75 Conference: Applications of Evolutionary Computing, EvoWorkshops 2006: EvoBIO, EvoCOMNET, EvoHOT, EvoIASP, EvoINTERACTION, EvoMUSART, and EvoSTOC, Budapest, Hungary, April 10-12, 2006, Proceedings
Source: DBLP

ABSTRACT The shaky ladder hyperplane-defined functions (sl-hdf's) are a test suite utilized for exploring the behavior of the genetic algorithm (GA) in dynamic environments. We present three ways of constructing the sl-hdf's by manipulating the way building blocks are constructed, combined, and changed. We examine the eect of the length of elemen- tary building blocks used to create higher building blocks, and the way in which those building blocks are combined. We show that the eects of building block construction on the behavior of the GA are complex. Our results suggest that construction routines which increase the roughness of the changes in the environment allow the GA to perform better by preventing premature convergence. Moreover, short length elementary building blocks permit early rapid progress.

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The shaky ladder hyperplane-defined functions (sl-hdfs) are a test suite utilized for exploring the behavior of the genetic algorithm (GA) in dynamic environments. This test suite can generate arbitrary problems with similar levels of difficulty and it provides a platform for systematic controlled observations of the GA in dynamic environments. Previous work has found two factors that contribute to the GA's success on sl-hdfs: (1) short initial building blocks and (2) significantly changing the reward structure during fitness landscape changes. Therefore a test function that combines these two features should facilitate even better GA performance. This has led to the construction of a new sl-hdf variant, "Defined Cliffs," in which we combine short elementary building blocks with sharp transitions in the environment. We examine this variant with two different levels of dynamics, static and regularly changing, using four different metrics. The results show superior GA performance on the Defined Cliffs over all previous variants (Cliffs, Weight, and Smooth). Our observations and conclusions in this variant further the understanding of the GA in dynamic environments.
    Genetic and Evolutionary Computation Conference, GECCO 2007, Proceedings, London, England, UK, July 7-11, 2007; 01/2007
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: One argument as to why the hyperplane-defined functions (hdf's) are a good testbed for the genetic algorithm (GA) is that the hdf's are built in the same way that the GA works. In this paper we test that hypothesis in a new setting by ex-ploring the GA on a subset of the hdf's which are dynamic, the shaky ladder hyperplane-defined functions (sl-hdf's). In doing so we gain insight into how the GA makes use of crossover during its traversal of the sl-hdf search space. We begin this paper by explaining the sl-hdf's. We then conduct a series of experiments with various crossover rates and var-ious rates of environmental change. Our results show that the GA performs better with than without crossover in dy-namic environments. Though these results have been shown on some static functions in the past, they are re-confirmed and expanded here for a new type of function (the hdf) and a new type of environment (dynamic environments). More-over we show that crossover is even more beneficial in dy-namic environments than it is in static environments. We discuss how these results can be used to develop a richer knowledge about the use of building blocks by the GA.
  • [Show abstract] [Hide abstract]
    ABSTRACT: Researchers examining genetic algorithms (GAs) in applied settings rarely have access to anything other than fitness values of the best individuals to observe the behavior of the GA. In particular, researchers do not know what schemata are present in the population. Even when researchers look beyond best fitness values, they concentrate on either performance related measures like average fitness and robustness, or low-level descriptions like bit-level diversity measures. To understand the behavior of the GA on dynamic problems, it would be useful to track what is occurring on the “semantic” level of schemata. Thus in this paper we examine the evolving “content” in terms of schemata, as the GA solves dynamic problems. This allows us to better understand the behavior of the GA in dynamic environments. We finish by summarizing this knowledge and speculate about future work to address some of the new problems that we discovered during these experiments.
    Applications of Evolutinary Computing, EvoWorkshops 2007: EvoCoMnet, EvoFIN, EvoIASP,EvoINTERACTION, EvoMUSART, EvoSTOC and EvoTransLog, Valencia, Spain, April11-13, 2007, Proceedings.; 01/2007

Full-text (3 Sources)

Available from
Sep 4, 2014