This paper uses the recent idea that the fundamental haploid-diploid lifecycle of eukaryotic organisms implements a rudimentary form of learning within evolution. A general approach for evolutionary computation is her...
详细信息
This paper uses the recent idea that the fundamental haploid-diploid lifecycle of eukaryotic organisms implements a rudimentary form of learning within evolution. A general approach for evolutionary computation is here derived that differs from all previous known work using diploid representations. The primary role of recombination is also changed from that previously considered in both natural and artificial evolution under the new view. Using well-known abstract tuneable models it is shown that varying fitness landscape ruggedness varies the benefit of the new approach.
Population size in evolutionary algorithms (EAs) is critical for their performance. In this paper, we first give a comprehensive review of existing population control methods. Then, a few representative methods are se...
详细信息
evolutionary algorithms (EAs) are randomized search heuristics and general-purpose solvers that mimic behavior seen in natural evolution, such as mutation or crossover. They are most often used, due to their flexibili...
详细信息
In this article, we discuss the progress achieved with the use of evolutionary algorithms for the analysis of H-1 Nuclear Magnetic Resonance spectra of solutes in orientationally ordered liquids. With these tools the ...
详细信息
In this article, we discuss the progress achieved with the use of evolutionary algorithms for the analysis of H-1 Nuclear Magnetic Resonance spectra of solutes in orientationally ordered liquids. With these tools the analysis of extremely complex spectra that were hitherto impossible to solve has now become eminently feasible. We discuss applications to 2 molecules of special interest: (a) hexamethylbenzene, which is a text book example of steric hindrance between adjacent rotating methyl groups;and (b) cyclohexane which is the standard example of interconversion between various molecular conformations. New interesting physics is obtained in both cases.
Various evolutionary algorithms are being developed to search the optimal solution of various problems in the real world. evolutionary algorithms search solutions showing the optimal fitness to given problem using the...
详细信息
Data generation has grown rapidly over the recent years. Different types of products and services are offered daily on the Internet. Finding out elegant, flexible and robust strategies to deal with this amount of data...
详细信息
ISBN:
(纸本)9781509060177
Data generation has grown rapidly over the recent years. Different types of products and services are offered daily on the Internet. Finding out elegant, flexible and robust strategies to deal with this amount of data in a static way is one goal of data mining, whilst the data stream mining works in dynamic environments. The searching of co-occurrence of items in data is a task of a data miming branch named association rule mining. The present paper investigates the use of evolutionary algorithms as well as artificial immune systems to extract association rules within item sets in both, static and dynamic, environments. We perform a number of experiments over datasets from the association rule mining literature, and compare their performances. A discussion in terms of computational time and measures of interest is made to conclude the proposed study.
The approximation error of an evolutionary algorithm is the fitness difference between the optimal solution and a solution found by the algorithm. In this paper, an initial error analysis has been made to evolutionary...
详细信息
ISBN:
(纸本)9781450349390
The approximation error of an evolutionary algorithm is the fitness difference between the optimal solution and a solution found by the algorithm. In this paper, an initial error analysis has been made to evolutionary algorithms for discrete optimization. First, the order of convergence and asymptotic error constant are defined. Then it is proven that for any EA, under particular initialization, its order of convergence is 1 and its asymptotic error constant equals to the spectral radius of the transition probability sub-matrix;if its transition probability sub-matrix is primitive or upper triangular with unique diagonal entries, then under random initialization, its order of convergence is 1 and its asymptotic error constant equals to the spectral radius of the transition probability sub-matrix. Our study reveals that evolutionary algorithms converge linearly to the optimal solution and the spectral radius of the transition probability sub-matrix is the main factor in affecting the approximation error.
In the last years, several real-world problems that require to optimise an increasing number of variables have appeared. This type of optimisation, called large-scale global optimisation, is hard due to the huge incre...
详细信息
In the last years, several real-world problems that require to optimise an increasing number of variables have appeared. This type of optimisation, called large-scale global optimisation, is hard due to the huge increase of the domain search due to the dimensionality. Large-scale global optimisation is a research area getting more attention in the last years, thus many algorithms, mainly evolutionary algorithms, have been specially designed to tackle it. In this paper, we give a brief introduction of several of them and their techniques, remarking techniques based on grouping of variables and memetic algorithms, because they are two promising approaches. Also, we have reviewed the winners of the different competitions in the area, to give a snapshot of the algorithms that have obtained the best results in this area. Finally, several interesting trends in the research area have been pointed out, and some future trends and challenges have been suggested.
Job Shop Scheduling Problem (JSSP) represents a real challenge for the researchers' community due to its complexity consisting in the plurality of resources that needs to be optimally used and the variety of goals...
详细信息
ISBN:
(纸本)9781509020478
Job Shop Scheduling Problem (JSSP) represents a real challenge for the researchers' community due to its complexity consisting in the plurality of resources that needs to be optimally used and the variety of goals that needs to be accomplished. This paper presents the implementation of three evolutionary algorithms (Genetic algorithms, Particle Swarm Optimization and Ant Colony Optimization) for the JSSP. The tests are made considered a set of classical benchmarks for the proposed problem and the obtained results are subject to comparison.
The resource constrained project scheduling problem(RCPSP) has received wide attention in the last 20 years with a number of evolutionary algorithms being proposed. Most of these algorithms can produce optimal or near...
详细信息
ISBN:
(纸本)9781509025978
The resource constrained project scheduling problem(RCPSP) has received wide attention in the last 20 years with a number of evolutionary algorithms being proposed. Most of these algorithms can produce optimal or near optimal solutions in less than a second. However, a close investigation of the literature will reveal a number of questionable benchmarking practices. In this paper I highlight some of these issues together with possible future research directions which are mainly centred around the use of hyper-heuristics.
暂无评论