Mark Wineberg

The underlying similarity of diversity measures used in evolutionary computation

In this paper we compare and analyze the various diversity measures used in the Evolutionary Computation field. While each measure looks quite different from the others in form, we surprisingly found that the same basic method underlies all of them: the distance between all possible pairs of chromosomes/ organisms in the population. This is true even of the Shannon entropy of gene frequencies. We then associate the different varieties of EC diversity measures to different diversity measures used in Biology. Finally we give an O(n) implementation for each of the diversity measures (where n is the population size), despite their basis in an O(n2) number of comparisons.

Distance between populations

Gene space, as it is currently formulated, cannot provide a solid basis for investigating the behavior of the GA. We instead propose an approach that takes population effects into account. Starting from a discussion of diversity, we develop a distance measure between populations and thereby a population metric space. We finally argue that one specific parameterization of this measure is particularly appropriate for use with GAs.

An Integrated Framework for a Joint Assimilation of Brightness Temperature and Soil Moisture Using the Nondominated Sorting Genetic Algorithm II

AbstractThis study has applied the Nondominated Sorting Genetic Algorithm II (NSGA-II) in a two-step assimilation procedure to jointly assimilate brightness temperature into a radiative transfer model and soil moisture into a land surface model. The first assimilation procedure generates a time series of soil moisture by assimilating brightness temperature from the Advanced Microwave Scanning Radiometer for Earth Observing System (AMSR-E) into the Land Parameter Retrieval Model (LPRM). The second procedure generates assimilated soil moisture by assimilating the soil moisture from LPRM into the Canadian Land Surface Scheme (CLASS). Note that the assimilated soil moisture was generated by merging two soil moisture estimates: one from LPRM and the other from the CLASS simulation. The assimilated soil moisture is better than using the soil moisture determined either from the satellite observation or the land surface scheme alone. This method provides improved model state and parameterizations for both LPRM an...

Estimation of evolvability genetic algorithm and dynamic environments

This article investigates the of applicability of adding evolvability promoting mechanisms to a genetic algorithm to enhance its ability to handle perpetually novel dynamic environments, especially one that has stationary periods allowing the Genetic Algorithm (GA) to converge on a temporary global optimum. We utilize both biological and evolutionary computation (EC) definitions of evolvability to create two measures: one based on the improvements in fitness; the other based on the amount of genotypic change. These two evolvability measures are used alongside fitness to modify how selection proceeds in the GA. We call this modified GA the Estimation of Evolvability Genetic Algorithm (EEGA). When tested against a regular GA (with random immigrants), the EEGA is able to track the global optimum more closely than the GA during the dynamic period. Unlike most GA extensions, the EEGA works effectively at a lower level of diversity than does the GA, showing that it is the quality of the diverse members in the population and not just the quantity that helps the GA evolve.

The Shifting Balance Genetic Algorithm as More than Just Another Island Model GA

The Shifting Balance Genetic Algorithm (SBGA) is an extension of the Genetic Algorithm (GA) that was created to promote guided diversity to improve performance in highly multimodal environments. In this paper a new behavioral model for the SBGA is presented. Based on the model, various modifications of the SBGA are proposed: these include a mechanism for managing dynamic population sizes along with population restarts. The various mechanisms that make up the SBGA are compared and contrasted against each other and against other Island Model GA systems. It was found that the mechanisms that characterize the SBGA, such as a repulsive central force from one population on the others, could improve the behavior of multi-populational systems.

An evolutionary approach to behavioural-level synthesis

This paper presents a novel approach to the concurrent solution of three high-level synthesis (HLS) problems and solves them in an integrated manner using hierarchical genetic algorithm (HGA). We focus on the core problems of HLS: scheduling, allocation, and binding. Scheduling consists of assigning of operations in an data-flow graph (DFG) to control steps or clock cycles. Allocation selects specific numbers and types of functional units from a hardware library to perform the operations specified in the DFG. Binding assigns constituent operations of the DFG to specific unit instances. A very general version of the problem is considered where functional units may perform different numbers of control steps. The HLS problems are solved by applying two genetic algorithms in a hierarchical manner. The first performs allocation, while the second performs scheduling and binding and serves as the fitness functions for the first. When compared to other, well-known techniques, our results show a reduction in time to obtain optimal solutions for standard benchmarks.

Enhancement of the shifting balance genetic algorithm for highly multimodal problems

The shifting balance genetic algorithm (SBGA) is an extension of the genetic algorithm (GA) that was created to promote guided diversity to improve performance in highly multimodal environments. Based on a new behavioral model for the SBGA, various modifications are proposed: these include a mechanism for managing dynamic population sizes with population restarts, and communication among the colonies. The enhanced SBGA is compared against the original SBGA system and other multipopulational GA systems on a complex mathematical function (F8F2) and on the NP-complete 0/1 knapsack problem. In all cases, the enhanced SBGA outperformed all other systems, and on the 0/1 knapsack problem, it was the only one to find the global optimum.

Depictions of genotypic space for evaluating the suitability of different recombination operators

When the genetic algorithm recombines two parent genotypes, the differences between them define a genotypic subspace, and any offspring produced should be confined to this subspace. Although this might seem insignificant, those recombination (or crossover) operators that violate this principle can direct a search away from the region (in genotypic space) that contains the two parent genotypes. This is contrary to the task for which the recombination operator was originally developed and can be detrimental, so this paper introduces a visualization that can be used to detect violations of this principle. The methodology also inspired the development of a different approach to recombining permutations, and a brief case study shows that an alternative recombination operator that does not violate this principle can be used to achieve a performance improvement over previous attempts to optimize Field-Programmable Gate-Array placements using a genetic algorithm. We believe that this technique will be invaluable for developing additional recombination operators.

Approaches to multidimensional scaling for adaptive landscape visualization

The adaptive landscape has become a standard approach for genetic algorithm visualization, and the representation of the higher dimensional chromosome space onto a two-dimensional plane suitable for the construction of an adaptive landscape requires an accurate measurement of the distance between chromosomes. Although the shortcomings of traditional approaches to adaptive landscape construction are by no means unknown to the research community, the intuitions afforded by this visualization have kept it in widespread usage. Since the multidimensional scaling required for the creation of a representative landscape is often disregarded to avoid the computational overhead required, this paper demonstrated that distance measures are available that remain representative of the genetic operators of the genetic algorithm while being suitable for multidimensional scaling techniques. This paper also demonstrated that in spite of the complications expected when the distance between chromosomes is measured with respect to both a unary mutation operation and a binary recombination operation simultaneously, it is possible to construct adaptive landscapes that depict features indicative of the effects of both genetic operators.

The estimation of evolvability genetic algorithm

In this paper we utilize both the biological and common EC definitions of evolvability to create two measures: one based on fitness improvement, the other based on the amount of genotypic change. The evolvability measures are then used to increase the exploratory behavior of the GA to escape from local optima and track moving environments. The estimation of evolvability genetic algorithm was successfully tested against the GA both in stationary and dynamic environments. The EEGA behaved so well that it was difficult to determine solely from the behavior of the EEGA when the function began moving. Furthermore, unlike most GA extensions created for dynamic environment, the EEGA actually performs at a lower diversity level than a standard GA.