In seeking constraints on global climate model projections under global warming, one commonly finds that different subsets of models perform well under different objective functions, and these trade-offs are difficult...
详细信息
In seeking constraints on global climate model projections under global warming, one commonly finds that different subsets of models perform well under different objective functions, and these trade-offs are difficult to weigh. Here a multiobjective approach is applied to a large set of subensembles generated from the Climate Model Intercomparison Project phase 5 ensemble. We use observations and reanalyses to constrain tropical Pacific sea surface temperatures, upper level zonal winds in the midlatitude Pacific, and California precipitation. An evolutionary algorithm identifies the set of Pareto-optimal subensembles across these three measures, and these subensembles are used to constrain end-of-century California wet season precipitation change. This methodology narrows the range of projections throughout California, increasing confidence in estimates of positive mean precipitation change. Finally, we show how this technique complements and generalizes emergent constraint approaches for restricting uncertainty in end-of-century projections within multimodel ensembles using multiple criteria for observational constraints.
This article proposes an alternative constraint handling technique for the four-bar linkage path generation problem. The constraint handling technique that is traditionally applied uses an exterior penalty function, a...
详细信息
This article proposes an alternative constraint handling technique for the four-bar linkage path generation problem. The constraint handling technique that is traditionally applied uses an exterior penalty function, and has been found to be inefficient, particularly when dealing with constraints on the input angles. The new technique deals with both input crank rotation and Grashof's criterion. Four classical path generation problems are used to test the performance of the proposed technique. A new adaptive teaching-learning-based optimization (TLBO) scheme is used to solve several optimization problems. This technique is referred to here as teaching-learning-based optimization with a diversity archive (ATLBO-DA), and was specifically developed for this design problem. The results show that this new design concept gives better results than those of previous work, and that ATLBO-DA is superior to the original version and other metaheuristic algorithms.
Over the last decade, deep neural networks have shown great success in the fields of machine learning and computer vision. Currently, the CNN (convolutional neural network) is one of the most successful networks, havi...
详细信息
Over the last decade, deep neural networks have shown great success in the fields of machine learning and computer vision. Currently, the CNN (convolutional neural network) is one of the most successful networks, having been applied in a wide variety of application domains, including pattern recognition, medical diagnosis and signal processing. Despite CNNs' impressive performance, their architectural design remains a significant challenge for researchers and practitioners. The problem of selecting hyperparameters is extremely important for these networks. The reason for this is that the search space grows exponentially in size as the number of layers increases. In fact, all existing classical and evolutionary pruning methods take as input an already pre-trained or designed architecture. None of them take pruning into account during the design process. However, to evaluate the quality and possible compactness of any generated architecture, filter pruning should be applied before the communication with the data set to compute the classification error. For instance, a medium-quality architecture in terms of classification could become a very light and accurate architecture after pruning, and vice versa. Many cases are possible, and the number of possibilities is huge. This motivated us to frame the whole process as a bi-level optimization problem where: (1) architecture generation is done at the upper level (with minimum NB and NNB) while (2) its filter pruning optimization is done at the lower level. Motivated by evolutionary algorithms' (EAs) success in bi-level optimization, we use the newly suggested co-evolutionary migration-based algorithm (CEMBA) as a search engine in this research to address our bi-level architectural optimization problem. The performance of our suggested technique, called Bi-CNN-D-C (Bi-level convolution neural network design and compression), is evaluated using the widely used benchmark data sets for image classification, called CIFAR-10, CI
This paper presents a simple and efficient real-coded genetic algorithm (RCGA) for constrained real-parameter optimization. Different from some conventional RCGAs that operate evolutionary operators in a series framew...
详细信息
This paper presents a simple and efficient real-coded genetic algorithm (RCGA) for constrained real-parameter optimization. Different from some conventional RCGAs that operate evolutionary operators in a series framework, the proposed RCGA implements three specially designed evolutionary operators, named the ranking selection (RS), direction-based crossover (DBX), and the dynamic random mutation (DRM), to mimic a specific evolutionary process that has a parallel-structured inner loop. A variety of benchmark constrained optimization problems (COPs) are used to evaluate the effectiveness and the applicability of the proposed RCGA. Besides, some existing state-of-the-art optimization algorithms in the same category of the proposed algorithm are considered and utilized as a rigorous base of performance evaluation. Extensive comparison results reveal that the proposed RCGA is superior to most of the comparison algorithms in providing a much faster convergence speed as well as a better solution accuracy, especially for problems subject to stringent equality constraints. Finally, as a specific application, the proposed RCGA is applied to optimize the GaAs film growth of a horizontal metal-organic chemical vapor deposition reactor. Simulation studies have confirmed the superior performance of the proposed RCGA in solving COPs. (C) 2015 Elsevier B.V. All rights reserved.
To select an adequate coding is one of the main problems in applications based on evolutionary algorithms. Many codings have been proposed to represent the search space for obtaining decision rules. A suitable represe...
详细信息
For any optimization algorithm tuning the parameters is necessary for effective and efficient optimization. We use a meta-level evolutionary algorithm for optimizing the effectiveness and efficiency of a load-balancin...
详细信息
The study conducted in this work analyses the interactions between different evolutionary algorithms when they are hybridized. For this purpose, the phylogenetic tree of the best solution reported by the hybrid algori...
详细信息
An evolutionary algorithm based on the parallel evolution of multiple single objective populations and Pareto archive population is proposed, which is not only suitable for solving multi-objective optimization, but al...
详细信息
The need for supporting CSCW applications with heterogeneous and varying user requirements calls for adaptive and reconfigurable schedulers accommodating a mixture of real-time, proportional share, fixed priority and ...
详细信息
In this paper, a novel Parallel Cellular Automata (PCA) approach is presented for multi-objective reservoir operation optimization. The problem considers the multi-objective operation of a single reservoir with the tw...
详细信息
In this paper, a novel Parallel Cellular Automata (PCA) approach is presented for multi-objective reservoir operation optimization. The problem considers the multi-objective operation of a single reservoir with the two conflicting objectives of water supply and energy production. The water supply objective is defined as the squared deviation of the monthly release from the downstream demand while the hydropower objective is defined as the squared deficit of the monthly power production from the installed capacity. The proposed method uses two parallel cellular automata methods each searching for the solution of a single objective problem starting from an initial random solution. Each CA, however, is randomly seeded with the solution provided by the other CA method at each CA iteration. Two different version of the proposed PCA is considered based on the way the CAs are seeded. In the first method referred to as PCA1, a fixed value of 0.5 is used for the probability of exchange while in the second method, referred to as PCA2, a temperature-based variable probability of exchange is used for seeding the CAs. The proposed methods are used for bi-objective operation of Dez reservoir in Iran. Various operation periods of 60, 120, 240 and 480 months are considered to illustrate the efficiency and effectiveness of the proposed PCA methods for problems of different scales. In addition, Non-dominated Sorting Genetic Algorithm (NSGAII), is also used to solve the problems and the results are presented and compared. The results indicate that Pareto solutions obtained by the proposed temperature based method PCA2 are well-scattered over the front and in particular toward the end points compared to those of NSGAII requiring much less computational time. The superiority of the proposed method to that of NSGAII is shown to increase with increasing scale of the problem.
暂无评论