You are here
Search results
(1 - 6 of 6)
- Title
- Achieving consistent evolution across isometrically equivalent search spaces
- Creator
- Patton, Arnold L.
- Date
- 2004
- Collection
- Electronic Theses & Dissertations
- Title
- Balancing convergence and diversity in evolutionary single, multi and many objectives
- Creator
- Seada, Haitham
- Date
- 2017
- Collection
- Electronic Theses & Dissertations
- Description
-
"Single objective optimization targets only one solution, that is usually the global optimum. On the other hand, the goal of multiobjective optimization is to represent the whole set of trade-off Pareto-optimal solutions to a problem. For over thirty years, researchers have been developing Evolutionary Multiobjective Optimization (EMO) algorithms for solving multiobjective optimization problems. Unfortunately, each of these algorithms were found to work well on a specific range of objective...
Show more"Single objective optimization targets only one solution, that is usually the global optimum. On the other hand, the goal of multiobjective optimization is to represent the whole set of trade-off Pareto-optimal solutions to a problem. For over thirty years, researchers have been developing Evolutionary Multiobjective Optimization (EMO) algorithms for solving multiobjective optimization problems. Unfortunately, each of these algorithms were found to work well on a specific range of objective dimensionality, i.e. number of objectives. Most researchers overlooked the idea of creating a cross-dimensional algorithm that can adapt its operation from one level of objective dimensionality to the other. One important aspect of creating such algorithm is achieving a careful balance between convergence and diversity. Researchers proposed several techniques aiming at dividing computational resources uniformly between these two goals. However, in many situations, only either of them is difficult to attain. Also for a new problem, it is difficult to tell beforehand if it will be challenging in terms of convergence, diversity or both. In this study, we propose several extensions to a state-of-the-art evolutionary many-objective optimization algorithm - NSGA-III. Our extensions collectively aim at (i) creating a unified optimization algorithm that dynamically adapts itself to single, multi- and many objectives, and (ii) enabling this algorithm to automatically focus on either convergence, diversity or both, according to the problem being considered. Our approach augments the already existing algorithm with a niching-based selection operator. It also utilizes the recently proposed Karush Kuhn Tucker Proximity Measure to identify ill-converged solutions, and finally, uses several combinations of point-to-point single objective local search procedures to remedy these solutions and enhance both convergence and diversity. Our extensions are shown to produce better results than state-of-the-art algorithms over a set of single, multi- and many-objective problems."--Pages ii-iii.
Show less
- Title
- Evolutionary optimization and ensemble techniques for data mining and pattern recognition
- Creator
- Topchy, Alexander P.
- Date
- 2004
- Collection
- Electronic Theses & Dissertations
- Title
- Natural niching : applying ecological principles to evolutionary computation
- Creator
- Goings, Sherri
- Date
- 2010
- Collection
- Electronic Theses & Dissertations
- Title
- The evolution of division of labor in digital organisms
- Creator
- Goldsby, Heather J.
- Date
- 2011
- Collection
- Electronic Theses & Dissertations
- Description
-
Division of labor is a hallmark strategy employed by a wide variety of groups ranging in complexity from bacteria to human economies. Within these groups, some individuals, such as worker ants, sacrifice their ability to reproduce and instead dedicate their lives to the maintenance of the colony and success of their kin. A worker ant may spend its entire life performing a single task, such as defending the colony or tending to the brood. The complexity of the strategies employed by these...
Show moreDivision of labor is a hallmark strategy employed by a wide variety of groups ranging in complexity from bacteria to human economies. Within these groups, some individuals, such as worker ants, sacrifice their ability to reproduce and instead dedicate their lives to the maintenance of the colony and success of their kin. A worker ant may spend its entire life performing a single task, such as defending the colony or tending to the brood. The complexity of the strategies employed by these groups, combined with their rampant success, gives rise to questions regarding why division of labor exists. While extensive research has been done to better understand the patterns and mechanisms of division of labor, exploring this topic in an evolutionary context remains challenging to study due to the slow pace of evolution and imperfect historical data. Understanding how and why division of labor arises is pertinent not just for understanding biological phenomena, but also as a means to enable evolutionary computation techniques to address complex problems using problem decomposition. The objective of problem-decomposition approaches is to have a group of individuals cooperatively solve a complex task by breaking it into pieces, having specialist individuals solve the pieces, and reassembling the solution. Essentially, problem-decomposition approaches use division of labor to enable groups to solve more challenging problems than any individual could alone. Unfortunately, human engineers have struggled with creating effective, automated problem-decomposition approaches.In this dissertation, I use digital evolution (i.e., populations of self-replicating computer programs that undergo open-ended evolution) to investigate questions related to the evolution of division of labor and to apply these insights to problem decomposition techniques. This dissertation has three primary components: First, we provide experimental evidence that evolutionary computation techniques can evolve groups of individuals that exhibit division of labor. Second, we explore two hypotheses for the evolution of division of labor. Specifically, we find support for the hypothesis that temporal polyethism (i.e., where a worker's age is related to the task it performs within the colony) may result from the evolutionary pressures of aging and risks associated with tasks. Additionally, we find support for a hypothesis initially proposed by Adam Smith, the premier economist, that the presence of task-switching costs results in an increase in the amount of division of labor exhibited by groups. Third, we describe how our analyses revealed that groups of organisms evolved as part of our task-switching work exhibit complex problem decomposition strategies that can potentially be applied to other evolutionary computation challenges. This work both informs biological studies of division of labor and also provides insights that can enable the development of new mechanisms for using evolutionary computation to solve increasingly complex engineering problems.
Show less
- Title
- Out of the box optimization using the parameter-less population pyramid
- Creator
- Goldman, Brian W.
- Date
- 2015
- Collection
- Electronic Theses & Dissertations
- Description
-
The Parameter-less Population Pyramid (P3) is a recently introduced method for performing evolutionary optimization without requiring any user-specified parameters. P3’s primary innovation is to replace the generational model with a pyramid of multiple populations that are iteratively created and expanded. In combination with local search and advanced crossover, P3 scales to problem difficulty, exploiting previously learned information before adding more diversity.Across seven problems, each...
Show moreThe Parameter-less Population Pyramid (P3) is a recently introduced method for performing evolutionary optimization without requiring any user-specified parameters. P3’s primary innovation is to replace the generational model with a pyramid of multiple populations that are iteratively created and expanded. In combination with local search and advanced crossover, P3 scales to problem difficulty, exploiting previously learned information before adding more diversity.Across seven problems, each tested using on average 18 problem sizes, P3 outperformed all five advanced comparison algorithms. This improvement includes requiring fewer evaluations to find the global optimum and better fitness when using the same number of evaluations. Using both algorithm analysis and comparison we show P3’s effectiveness is due to its ability to properly maintain, add, and exploit diversity. Unlike the best comparison algorithms, P3 was able to achieve this quality without any problem-specific tuning. Thus, unlike previous parameter-less methods, P3 does not sacrifice quality for applicability. Therefore we conclude that P3 is an efficient, general, parameter-less approach to black-box optimization that is more effective than existing state-of-the-art techniques.Furthermore, P3 can be specialized for gray-box problems, which have known, limited, non-linear relationships between variables. Gray-Box P3 leverages the Hamming-Ball Hill Climber, an exceptionally efficient form of local search, as well as a novel method for performing crossover using the known variable interactions. In doing so Gray-Box P3 is able to find the global optimum of large problems in seconds, improving over Black-Box P3 by up to two orders of magnitude.
Show less