Showing papers on "Genetic algorithm published in 1992"
TL;DR: In this paper, the loss minimum reconfiguration problem in the open loop radial distribution system is formulated as a mixed integer programming problem and a detailed solution methodology by the use of genetic algorithm is outlined.
Abstract: The loss minimum reconfiguration problem in the open loop radial distribution system is basically one of complex combinatorial optimization, since the normal open sectionalizing switches must be determined appropriately. The genetic algorithm was successfully applied to the loss minimum reconfiguration problem. In the proposed algorithm, strings consist of sectionalizing switch status or radial configurations, and the fitness function consists of the total system losses and penalty value of voltage drop and current capacity violations. The loss minimum reconfiguration problem is formulated as a mixed integer programming problem. The essential components of the genetic algorithm are briefly described. A detailed solution methodology by the use of genetic algorithm is outlined. Numerical examples demonstrate the validity and effectiveness of the proposed methodology. >
700 citations
Proceedings Article•
01 Jan 1992TL;DR: This paper presents a modification of the standard generational genetic algorithm that is designed to maintain the diversity required to track a changing response surface and shows some promise for the new technique.
Abstract: Genetic algorithms perform an adaptive search by maintaining a population of candidate solutions that are allocated dynamically to promising regions of the search space. The distributed nature of the genetic search provides a natural source of power for searching in changing environments. As long as sufficient diversity remains in the population the genetic algorithm can respond to a changing response surface by reallocating future trials. However, the tendency of genetic algorithms to converge rapidly reduces their ability to identify regions of the search space that might suddenly become more attractive as the environment changes. This paper presents a modification of the standard generational genetic algorithm that is designed to maintain the diversity required to track a changing response surface. An experimental study shows some promise for the new technique.
615 citations
TL;DR: In this paper, the use of a genetic algorithm to optimize the stacking sequence of a composite laminate for buckling load maximization is studied, and the advantage of the genetic algorithm in producing several near-optimal designs is discussed.
Abstract: The use of a genetic algorithm to optimize the stacking sequence of a composite laminate for buckling load maximization is studied. Various genetic parameters including the population size, the probability of mutation, and the probability of crossover are optimized by numerical experiments. A new genetic operator - permutation - is proposed and shown to be effective in reducing the cost of the genetic search. Results are obtained for a graphite-epoxy plate, first when only the buckling load is considered, and then when constraints on ply contiguity and strain failure are added. The influence on the genetic search of the penalty parameter enforcing the contiguity constraint is studied. The advantage of the genetic algorithm in producing several near-optimal designs is discussed.
576 citations
TL;DR: For folding on a simple two-dimensional lattice it is found that the genetic algorithm is dramatically superior to conventional Monte Carlo methods.
545 citations
TL;DR: An overview of the field of genetic algorithms is presented as representing a novel optimization strategy which is receiving much attention and is based on the principles of population genetics and biology.
Abstract: Presents an overview of the field of genetic algorithms, pioneered in the field of natural adaptive systems and simulated in software. They are shown as representing a novel optimization strategy which is receiving much attention. In machine learning they are a component of classifier systems which are able to extract rules from data. The algorithms discussed are based on the principles of population genetics and biology.
373 citations
12 May 1992
TL;DR: An efficient genetic algorithm for two NP-hard problems, the bin packing and the line balancing problems, and an encoding of solutions of fitting these problems is presented.
Abstract: The authors present an efficient genetic algorithm for two NP-hard problems, the bin packing and the line balancing problems. They define the two problems precisely and specify a cost function suitable for the bin packing problem. It is shown that the classic genetic algorithm performs poorly on grouping problems and an encoding of solutions of fitting these problems is presented. Efficient crossover and mutation operators are introduced for bin packing. The modification necessary to fit these operators for line balancing is given. Results of performance tests on randomly generated data are included. The line balancing tests cover real-world problem sizes. The results and areas of further research are discussed. >
296 citations
Proceedings Article•
01 Jan 1992
287 citations
Proceedings Article•
01 Jan 1992
TL;DR: Stabilization is enhanced by combining with other known stabilizers, while the use of the alkyl nitrates or alkynols eliminates the need for a nitroalkane in the stabilizer formulation.
Abstract: An alkyl alkynyl sulfide can be employed to stabilizer methylchloroform against reaction with the common metals of construction. Stabilization is enhanced by combining with other known stabilizers. Nitroalkanes, alkyl nitrates or alkynols may be employed to eliminate the need of dioxane and alkylene oxides, while the use of the alkyl nitrates or alkynols eliminates the need for a nitroalkane in the stabilizer formulation.
281 citations
TL;DR: The genetic algorithm technique is used to design a lateral autopilot and a windshear controller and shows that a variety of aerospace control system optimization problems can be addressed using genetic algorithms with no special problem-dependent modifications.
Abstract: The use of genetic algorithms as a technique for solving aerospace-related control system optimization problems is explored in this paper. Genetic algorithms are parameter search procedures based on the mechanics of natural genetics. They combine a Darwinian survival-of-the-fittest strategy with a random yet structured information exchange among a population of artificial chromosomes. The genetic algorithm technique is used to design a lateral autopilot and a windshear controller. The results show that a variety of aerospace control system optimization problems can be addressed using genetic algorithms with no special problem-dependent modifications. Suggestions for other uses related to aerospace control system optimization are presented.
278 citations
TL;DR: In this paper, a bi-criteria mathematical model with a solution procedure based on a genetic algorithm is proposed for the formation of machine cells and component families in a cellular manufacturing environment.
273 citations
TL;DR: Numerical results obtained here are compared with ones yielded by GAMS, a system for construction and solution of large and complex mathematical programming models, which appears to work well only for linear quadratic optimal control problems or problems with short horizon.
Abstract: This paper studies the application of a genetic algorithm to discrete-time optimal control problems. Numerical results obtained here are compared with ones yielded by GAMS, a system for construction and solution of large and complex mathematical programming models. While GAMS appears to work well only for linear quadratic optimal control problems or problems with short horizon, the genetic algorithm applies to more general problems equally well.
10 Nov 1992
TL;DR: An approach being explored to improve the usefulness of machine learning techniques for generating classification rules for complex, real-world data is described, which involves the use of genetic algorithms as a front end to a traditional rule induction system.
Abstract: An approach being explored to improve the usefulness of machine learning techniques for generating classification rules for complex, real-world data is described. The approach involves the use of genetic algorithms as a front end to a traditional rule induction system in order to identify and select the best subset of features to be used by the rule induction system. This approach has been implemented and tested on difficult texture classification problems. The results are encouraging and indicate that there are significant advantages to the approach in this domain. >
TL;DR: The authors describe experiments using a genetic algorithm for feature selection in the context of neural network classifiers, specifically, counterpropagation networks, and propose a method called the training set sampling, which selects feature sets that are as good as and occasionally better for counterPropagation than those chosen by an evaluation that uses the entire training set.
Abstract: The authors describe experiments using a genetic algorithm for feature selection in the context of neural network classifiers, specifically, counterpropagation networks. They present the novel techniques used in the application of genetic algorithms. First, the genetic algorithm is configured to use an approximate evaluation in order to reduce significantly the computation required. In particular, though the desired classifiers are counterpropagation networks, they use a nearest-neighbor classifier to evaluate features sets and show that the features selected by this method are effective in the context of counterpropagation networks. Second, a method called the training set sampling in which only a portion of the training set is used on any given evaluation, is proposed. Computational savings can be made using this method, i.e., evaluations can be made over an order of magnitude faster. This method selects feature sets that are as good as and occasionally better for counterpropagation than those chosen by an evaluation that uses the entire training set. >
TL;DR: Dynamic Parameter Encoding is shown to be empirically effective and amenable to analysis; the problem of premature convergence in GAs is explored through two convergence models.
Abstract: The common use of static binary place-value codes for real-valued parameters of the phenotype in Holland's genetic algorithm (GA) forces either the sacrifice of representational precision for efficiency of search or vice versa. Dynamic Parameter Encoding (DPE) is a mechanism that avoids this dilemma by using convergence statistics derived from the GA population to adaptively control the mapping from fixed-length binary genes to real values. DPE is shown to be empirically effective and amenable to analysiss we explore the problem of premature convergence in GAs through two convergence models.
TL;DR: In this paper, a coding scheme designed for facility layout is presented, where a solution is represented by the postorder sequence of the nodes in a slicing tree, and new solutions are generated by applying the various genetic operators in each generation.
12 May 1992
TL;DR: The authors describe the staged evolution of a complex motor pattern generator (MPG) for the control of a walking robot that was composed of a network of neurons with weights determined by genetic algorithm optimization.
Abstract: The authors describe the staged evolution of a complex motor pattern generator (MPG) for the control of a walking robot. The experiments were carried out on a six-legged, Brooks-style insect robot. The MPG was composed of a network of neurons with weights determined by genetic algorithm optimization. Staged evolution was used to improve the convergence rate of the algorithm. First, an oscillator for the individual leg movements was evolved. Then, a network of these oscillators was evolved to coordinate the movements of the different legs. By introducing a staged set of manageable challenges, the algorithm's performance was improved. >
TL;DR: In this article, an implementation of genetic search methods in the optimal design of structural systems with a mix of continuous, integer and discrete design variables is described, and the performance of each is evaluated in the context of structural design problems.
Abstract: The paper describes an implementation of genetic search methods in the optimal design of structural systems with a mix of continuous, integer and discrete design variables. Design variable representation schemes for such mixed variables are proposed and the performance of each is evaluated in the context of structural design problems. The approach is proposed as an alternative to the branch-and-bound techniques that are used in conjunction with nonlinear programming methods. The methodology is inherently equipped with a better chance of locating the global optimum than the conventional gradient based methods.
06 Jun 1992
TL;DR: The empirical studies show that the SGA can efficiently determine the network size and topology along with the optimal set of connection weights appropriate for desired tasks, without using backpropagation or any other learning algorithm.
Abstract: Presents a different type of genetic algorithm called the structured genetic algorithm (SGA) for the design of application-specific neural networks. The novelty of this new genetic approach is that it can determine the network structures and their weights solely by an evolutionary process. This is made possible for the SGA primarily due to its redundant genetic material and a gene activation mechanism which in combination provide a multi-layered structure to the chromosome. The authors focus on the use of this learning algorithm for automatic generation of a complete application specific neural network. With this approach, no a priori assumptions about topology are needed and the only information required is the input and output characteristics of the task. The empirical studies show that the SGA can efficiently determine the network size and topology along with the optimal set of connection weights appropriate for desired tasks, without using backpropagation or any other learning algorithm. >
Proceedings Article•
01 Jan 1992
TL;DR: Important aspects of sGA are presented which are able to exploit the repeatability of many nonstationary function optimization problems and suggest that sGA can solve complex problems more eeciently than has been possible with simple GAs.
Abstract: In this paper, we describe the application of a new type of genetic algorithm called the Structured Genetic Algorithm (sGA) for function optimization in nonstationary environments. The novelty of this genetic model lies primarily in its redundant genetic material and a gene activation mechanism which utilizes a multi-layered structure for the chromosome. In adapting to nonstationary environments of a repeated nature genes of long-term utility can be retained for rapid future deployment when favourable environments recur. The additional genetic material preserves optional solution space and works as a long term distributed memory within the population structure. This paper presents important aspects of sGA which are able to exploit the repeatability of many nonstation-ary function optimization problems. Theoretical arguments and empirical study suggest that sGA can solve complex problems more eeciently than has been possible with simple GAs. We also noted that sGA exhibits implicit genetic diversity and viability as in biological systems.
TL;DR: The author presents a learning algorithm and capabilities of perceptron-like neural networks whose outputs and inputs are directly connected to plants just like ordinary feedback controllers, andSimulations demonstrate that these networks can be optimized in terms of various evaluations, and they can discover schemes by themselves, such as state estimation and nonlinear control.
Abstract: The author presents a learning algorithm and capabilities of perceptron-like neural networks whose outputs and inputs are directly connected to plants just like ordinary feedback controllers. This simple configuration includes the difficulty of teaching the network. In addition, it is preferable to let the network learn so that a global and arbitrary evaluation of the total responses of the plant will be optimized eventually. In order to satisfy these needs, genetic algorithms are modified to accommodate the network learning procedure. This procedure is a kind of simulated evolution process in which a group of networks gradually improves as a whole, by crossing over connection weights among them, or by mutational changes of the weights, according to fitness values assigned to each network by a global evaluation. Simulations demonstrate that these networks can be optimized in terms of various evaluations, and they can discover schemes by themselves, such as state estimation and nonlinear control. >
Journal Article•
TL;DR: The implementation of a genetic algorithm (GA) to produce optimal or near-optimal intersection traffic signal timing strategies is described, with a focus on examining this application within a simple traffic situation, giving the reader a clear understanding of how the genetic algorithm is used.
Abstract: The implementation of a genetic algorithm (GA) (an artificial intelligence technique) to produce optimal or near-optimal intersection traffic signal timing strategies is described. The focus is on examining this application within a simple traffic situation, giving the reader a clear understanding of how the genetic algorithm is used. The problem involves finding a signal timing strategy that produces the smoothest traffic flow with the least average automobile delay. The problem domain has many tentative solutions. Therefore, signal timing design is expected to benefit from the parallel, global, and robust search characteristics of GAs. This gain is realized on a simulated four-intersection traffic network in the current implementation. The GA, by considering how traffic moves among multiple intersections (through simulation), can find a logical, near-optimal timing configuration. When this timing configuration is used in the corresponding real-world traffic situation, minimal total automobile delay is expected.
TL;DR: A learning algorithm for neural networks based on genetic algorithms is proposed and a simplified model for a brain with sensory and motor neurons is studied, whose structure is solely determined by an evolutionary process.
01 Jan 1992
TL;DR: Two models of the simple genetic algorithm are reviewed, extended, and unified and the result incorporates both short term (transient) and long term (asymptotic) GA behavior, leading to a geometric interpretation of genetic search which explains population trajectories.
Abstract: Two models of the simple genetic algorithm are reviewed, extended, and unified. The result incorporates both short term (transient) and long term (asymptotic) GA behavior. This leads to a geometric interpretation of genetic search which explains population trajectories.
TL;DR: In this article, the use of genetic algorithms (GAs) for the design of composite laminates is presented, where the design variables are the lamina orientations and stacking sequence required for maximum laminate strength and/or stiffness with minimum weight.
TL;DR: This article develops the thesis that parameter setting is correctly viewed as nondeductive, and discusses three basic learnability properties that must characterize the learner's linguistic environment.
Abstract: Most recent approaches to language learnability and acquisition have assumed that parameter setting is largely a deductive process. This article develops the thesis that parameter setting is correctly viewed as nondeductive. In particular, deductive approaches can be computationally costly and, in the worst case, are equal in cost to a brute enumerative search through the hypothesis space. The approach developed here uses natural selection, as simulated by a genetic algorithm, to simulate: parameter setting. A method is developed for evaluating the behavior of parsing devices relative to an environment (the input text), translating between parsing devices and a genome (a hypothesis string), and combining hypotheses via mating and mutation. A learner based on such a system will eventually arrive at the grammar for the least language compatible with its environment. We discuss three basic learnability properties that must characterize the learner's linguistic environment. Finally, we develop some recommenda...
TL;DR: A computational environment suitable for optimum design of structures in the general class of plane frames is described and the use of the environment is illustrated in a study of a cable‐stayed bridge structure.
Abstract: A computational environment suitable for optimum design of structures in the general class of plane frames is described. Design optimization is based on the use of a genetic algorithm in which a population of individual designs is changed generation by generation applying principles of natural selection and survival of the fittest. The fitness of a design is assessed using an objective function in which violations of design constraints are penalized. Facilities are provided for automatic data editing and reanalysis of the structure. The environment is particularly useful when parametric studies are required. The use of the environment is illustrated in a study of a cable‐stayed bridge structure.
06 Jun 1992
TL;DR: The GA is shown to be able to resolve the permutations, so that the advantages of an increase in the number of maxima outweigh the difficulties of recombination.
Abstract: The specification of neural net architectures by genetic algorithm (GA) is thought to be hampered by difficulties with crossover. This is the 'permutation' or 'competing conventions' problem: similar nets may have the hidden units defined in different orders so that they have very dissimilar genetic strings, preventing successful recombination of building blocks. Previous empirical tests of a number of recombination operators using a simulated net-building task indicated the superiority of one that sorts hidden unit definitions by overlap prior to crossover. However, simple crossover also fared well, suggesting that the permutation problem is not serious in practice. This is supported by an observed reduction in performance when the permutation problem is removed. The GA is shown to be able to resolve the permutations, so that the advantages of an increase in the number of maxima outweigh the difficulties of recombination. >
TL;DR: In this article, the authors used simulated annealing and genetic algorithms to find a one-dimensional earth structure which produces a seismogram that agrees with an observed seismogram in a geophysical inverse problem.
01 Jan 1992
TL;DR: The power of the parallel genetic algorithm (PGA) is shown with two combinatorial problems - the traveling salesman problem and the m graph partitioning problem, where the PGA has found solutions of very large problems, which are comparable or even better than any other solution found by other heuristics.
Abstract: Random search methods based on biological principles have been already proposed in the 60's. Our parallel genetic algorithm (PGA) uses two major modifications compared to the genetic algorithm. Firstly, selection for mating is distributed. Individuals live in a 2-D world. Selection of a mate is done by each individual independently in its neighborhood. Secondly, each individual may improve its fitness during its lifetime by e.g. local hill-climbing. The PGA runs with maximal efficiency on MIMD parallel computers. The search strategy of the PGA is based on a small number of intelligent and active individuals, whereas a GA uses a large population of passive individuals. We will show the power of the PGA with two combinatorial problems - the traveling salesman problem and the m graph partitioning problem. In these examples, the PGA has found solutions of very large problems, which are comparable or even better than any other solution found by other heuristics. A comparison between the PGA search strategy and iterated local hill-climbing is made.
TL;DR: In this article, an enhanced version of the genetic algorithm, using the notions of diploidy and dominance, was used to address the robustness issue of a fixed plant.
Abstract: An important aspect of a good control algorithm is robustness with respect to changing plant parameters. Adaptive control strategies attempt to address the robustness issue by optimizing control parameters as changes occur in the plant. This paper investigates the genetic algorithm as one possible means of adaptively optimizing the gains of a proportional-plus-integral hydrogenerator governor. Previous work has shown that the genetic algorithm can effectively optimize the control parameters for a fixed plant. An enhanced version of the genetic algorithm, using the notions of diploidy and dominance, can be used to address the robustness issue. Here, changes in the conduit time constant as well as load constant are considered. It is shown that the genetic algorithm can effectively follow changes in the plant parameters, producing optimal control parameters in an adaptive environment. >