This paper describes the N-Tuple Bandit Evolutionary algorithm (NTBEA), an optimisation algorithm developed for noisy and expensive discrete (combinatorial) optimisation problems. The algorithm is applied to two game-...
详细信息
ISBN:
(纸本)9781509060177
This paper describes the N-Tuple Bandit Evolutionary algorithm (NTBEA), an optimisation algorithm developed for noisy and expensive discrete (combinatorial) optimisation problems. The algorithm is applied to two game-based hyperparameter optimisation problems. The N-Tuple system directly models the statistics, approximating the fitness and number of evaluations of each modelled combination of parameters. The model is simple, efficient and informative. Results show that the NTBEA significantly outperforms grid search and an estimation of distribution algorithm.
We consider Black-Box continuous optimization by estimation of distribution algorithms (EDA). In continuous EDA, the multivariate Gaussian distribution is widely used as a search operator, and it has the well-known ad...
详细信息
ISBN:
(纸本)9783319108407;9783319108391
We consider Black-Box continuous optimization by estimation of distribution algorithms (EDA). In continuous EDA, the multivariate Gaussian distribution is widely used as a search operator, and it has the well-known advantage of modelling the correlation structure of the search variables, which univariate EDA lacks. However, the Gaussian distribution as a search operator is prone to premature convergence when the population is far from the optimum. Recent work suggests that replacing the univariate Gaussian with a univariate Cauchy distribution in EDA holds promise in alleviating this problem because it is able to make larger jumps in the search space due to the Cauchy distribution's heavy tails. In this paper, we propose the use of a multivariate Cauchy distribution to blend together the advantages of multivariate modelling with the ability of escaping early convergence to efficiently explore the search space. Experiments on 16 benchmark functions demonstrate the superiority of multivariate Cauchy EDA against univariate Cauchy EDA, and its advantages against multivariate Gaussian EDA when the population lies far from the optimum.
More and more experimental results have shown that genes do not function independently;instead they act on each other. To find the interactions among genes is one of the hottest topics in current genome research. The ...
详细信息
ISBN:
(纸本)9812565329
More and more experimental results have shown that genes do not function independently;instead they act on each other. To find the interactions among genes is one of the hottest topics in current genome research. The Bayesian Network (BN), which is a graph-based representation of a joint probability distribution that captures properties of conditional independence between variables, is a desirable tool to find the interaction between genes. However, how to find appropriate BNs that most fit to the data is very difficult since the number of possible BNs on n variables is the super-exponential of n. To avert the combinational explosion, in this paper, we use estimation of distribution algorithm (EDA) to search the BN space. Also, in order to make the individuals of EDA meaningful, we also propose a depth-first search method to cut circles in the graph. We have tested our method on cell-cycle gene expression data, the results show that the constructed BNs can not only discover some existing relationships in other literatures and Gene Ontology, but also reveal some previously unknown interactions.
To avoid blackouts in the energy system, the knowledge of the state of the power lines has critical importance. To get the timely conscience, Phasor Measurement Units (PMUs) are used to provide real-time synchronized ...
详细信息
ISBN:
(纸本)9781467368131
To avoid blackouts in the energy system, the knowledge of the state of the power lines has critical importance. To get the timely conscience, Phasor Measurement Units (PMUs) are used to provide real-time synchronized measurements of voltage and current phasors of the buses. The deployment of PMUs is done to first detect any single or multiple line outage, and further to identify the correct line in outage. But if these approaches are applied to more multiple line outages, the search space grows exponentially. To reduce the computational complexity in case of multiple line outage detection, stochastic optimization methods can be used. These methods give near optimal solution with an acceptable computational complexity. In this paper, the use of phasor angle measurements and optimal deployment of PMUs to identify multiple power line outages is investigated. An adaptive version of estimation of distribution algorithm (EDA) is proposed to detect and identify the lines in outage. In this adaptive EDA thresholding technique is introduced in order to get better solutions. It is shown that the proposed algorithm is achieving a better success rate than the other evolutionary techniques.
The large-scale transmission network expansion planning requires high computational speed and accuracy, so combining the DE (differential evolution) algorithm with the EDA (estimation of distribution algorithm), whose...
详细信息
ISBN:
(纸本)9781538649503
The large-scale transmission network expansion planning requires high computational speed and accuracy, so combining the DE (differential evolution) algorithm with the EDA (estimation of distribution algorithm), whose algorithm probability update mechanism is improved according to the characteristics of the transmission network expansion problem, a EDA/DE hybrid algorithm is presented to solve large-scale transmission network expansion planning problem. At the same time, taking into account the grid company investments, incentive policy for renewable energy as well as the security constraints, based on the embedded cost, wind curtailment and risk value, a multi-objective static planning model for transmission network expansion and wind power network optimization is established. IEEE24 RTS node example demonstrates the validity and applicability of the proposed algorithm.
This paper studies the probabilistic based evolutionary algorithms in dealing with bi-objective travelling salesman problem. Multi-objective restricted Boltzmann machine and univariate marginal distributionalgorithm ...
详细信息
ISBN:
(纸本)9783642172977
This paper studies the probabilistic based evolutionary algorithms in dealing with bi-objective travelling salesman problem. Multi-objective restricted Boltzmann machine and univariate marginal distributionalgorithm in binary representation are modified into permutation based representation. Each city is represented by an integer number and the probability distributions of the cities are constructed by running the modeling approach. A refinement operator and a local exploitation operator are proposed in this work. The probabilistic based evolutionary optimizers are subsequently combined with genetic based evolutionary optimizer to complement the limitations of both algorithms.
This paper adapts parallel master-slave estimation of distribution and genetic algorithms (GAs and EDAs) hybridization. The master selects portions of the search space, and slaves perform, in parallel and independentl...
详细信息
ISBN:
(纸本)9780769548937;9781467346245
This paper adapts parallel master-slave estimation of distribution and genetic algorithms (GAs and EDAs) hybridization. The master selects portions of the search space, and slaves perform, in parallel and independently, a GA that solves the problem on the assigned portion of the search space. The master's work is to progressively narrow the areas explored by the slave's GAs, using parallel dynamic K-means clustering to determine the basins of attraction of the search space. Coordination of activities between master and slaves is done in an asynchronous way (i.e. no waiting is entertained among the processes). The proposed asynchronous model has managed to reduce computation time while maintaining the quality of solutions.
Genetic & Evolutionary Biometrics (GEB) is a newly emerging area of study devoted to the design, analysis, and application of genetic and evolutionary computing to the field of biometrics. In this paper, we presen...
详细信息
ISBN:
(纸本)9781467313759
Genetic & Evolutionary Biometrics (GEB) is a newly emerging area of study devoted to the design, analysis, and application of genetic and evolutionary computing to the field of biometrics. In this paper, we present a GEB application called GEFEML (Genetic and Evolutionary Feature Extraction Machine Learning). GEFEML incorporates a machine learning technique, referred to as cross validation, in an effort to evolve a population of local binary pattern feature extractors (FEs) that generalize well to unseen subjects. GEFEML was trained on a dataset taken from the FRGC database and generalized well on two test sets of unseen subjects taken from the FRGC and MORPH databases. GEFEML evolved FEs that used fewer patches, had comparable accuracy, and were 54% less expensive in terms of computational complexity.
The Quadratic Assignment Problem (QAP) is a specially challenging permutation-based np-hard combinatorial optimization problem, since instances of size n > 40 are seldom solved using exact methods. In this sense, m...
详细信息
ISBN:
(纸本)9781450367486
The Quadratic Assignment Problem (QAP) is a specially challenging permutation-based np-hard combinatorial optimization problem, since instances of size n > 40 are seldom solved using exact methods. In this sense, many approximate methods have been published to tackle this problem, including estimation of distribution algorithms (EDAs). In particular, EDAs have been used to solve permutation problems by introducing distance based exponential models, such as the Mallows Models. In this paper we approximate the QAP with a Hamming distance based kernels of Mallows Models. Based on the benchmark instances, we have observed that our approach is competitive, reaching the best-known solution in 71% of the tested instances, especially on large instances (n > 125), where it is able to outperform state of the art results in 43 out of 288 instances.
Hybrid algorithms incorporated with parallel processing techniques are very powerful tools for efficiently solving very complex optimization problems. We present asynchronous parallel computer architecture adaptation ...
详细信息
ISBN:
(纸本)9781479941735
Hybrid algorithms incorporated with parallel processing techniques are very powerful tools for efficiently solving very complex optimization problems. We present asynchronous parallel computer architecture adaptation based on hybridization of Genetic algorithms (GAs) and estimation of distribution algorithms (EDAs). In this master-slave formulation, slaves perform evolutionary computation independently using GAs, while master supervises and controls the searching process. Master's role is to probabilistically study the characteristics of solution space and directs the slaves on good searching spots. This study reports some few findings on the ability of our hybrid algorithm to solve some instances of BQP problem as well as AODV routing optimization in VANETs. For both problems our hybrid algorithm has obtained best results in terms of quality of solutions as well as computational speed.
暂无评论