In the paper, particle gradient multi-objective evolutionary algorithm (PGMOEA) on GPU is presented. PGMOEA extends the classical particle dynamic multiobjective evolutionary algorithm by incorporating the gradient in...
详细信息
In the paper, particle gradient multi-objective evolutionary algorithm (PGMOEA) on GPU is presented. PGMOEA extends the classical particle dynamic multiobjective evolutionary algorithm by incorporating the gradient information of each particle from evolutionary programming. We perform experiments to compare PGMOEA on GPU with PGMOEA on CPU and demonstrate that PGMOEA on GPU is much more effective and efficient than PGMOEA on CPU.
In a wireless sensor network (WSN), the unbalanced distribution of communication loads often causes the problem of energy hole, which means the energy of the nodes in the hole region will be exhausted sooner than the ...
详细信息
In a wireless sensor network (WSN), the unbalanced distribution of communication loads often causes the problem of energy hole, which means the energy of the nodes in the hole region will be exhausted sooner than the nodes in other regions. This is a key factor which affects the lifetime of the networks. In this paper we propose an improved corona model with levels for analyzing sensors with adjustable transmission ranges in a WSN with circular multi-hop deployment (modeled as concentric coronas). Based on the model we consider that the right transmission ranges of sensors in each corona is the decision factor for optimizing the network lifetime after nodes deployment. We prove that searching optimal transmission ranges of sensors among all coronas is a multi-objective optimization problem (MOP). which is NP hard. Therefore, we propose a centralized algorithm and a distributed algorithm for assigning the transmission ranges of sensors in each corona for different node distributions. The two algorithms can not only reduce the searching complexity but also obtain results approximated to the optimal solution. Furthermore, the simulation results of our solutions indicate that the network lifetime approximates to that ensured by the optimal under both uniform and non-uniform node distribution. (C) 2009 Elsevier B.V. All rights reserved.
The problem of optimizing an absorber system for three-dimensional seismic structures is addressed. The objective is to determine the number and position of absorbers to minimize the coupling effects of translation-to...
详细信息
ISBN:
(纸本)9780819475480
The problem of optimizing an absorber system for three-dimensional seismic structures is addressed. The objective is to determine the number and position of absorbers to minimize the coupling effects of translation-torsion of structures at minimum cost. A procedure for a multi-objective optimization problem is developed by integrating a dominance-based selection operator and a dominance-based penalty function method. Based on the two-branch tournament genetic algorithm, the selection operator is constructed by evaluating individuals according to their dominance in one run. The technique guarantees the better performing individual winning its competition, provides a slight selection pressure toward individuals and maintains diversity in the population. Moreover, due to the evaluation for individuals in each generation being finished in one run, less computational effort is taken. Penalty function methods are generally used to transform a constrained optimizationproblem into an unconstrained one. The dominance-based penalty function contains necessary information on non-dominated character and infeasible position of an individual, essential for success in seeking a Pareto optimal set. The proposed approach is used to obtain a set of non-dominated designs for a six-storey three-dimensional building with shape memory alloy dampers subjected to earthquake.
With the development of embedded system and microelectronics technology, the Hardware/Software Co-Design Methodology of SoC has taken an important role in designing of embedded applications. Hardware/Software partitio...
详细信息
ISBN:
(纸本)9780769538167
With the development of embedded system and microelectronics technology, the Hardware/Software Co-Design Methodology of SoC has taken an important role in designing of embedded applications. Hardware/Software partitioning is one of the most significant part of Hardware/Software co-design of embedded systems. Hardware/Software partitioning is essentially multiobjectiveoptimizationproblem, namely, how to obtain an optimal hardware/software combination based on satisfying system function requirements and constraints. This paper presents an Immune Algorithm based on the Pareto concept of multi-objective optimization problems. Compared with other algorithm, it can provide an effective tool for measuring the performance of different objective functions, and improve the designing efficiency. In the last, the result was compared with those of using the single-objectiveoptimization on different task graphs. Experimental results show the algorithm can achieve the global optimal solution of the HW/SW partitioning problem based on meet system constraints.
multi-objective genetic algorithm is proved to be suitable for solving multi-objective optimization problems. However, it is usually very hard to balance the convergence and diversity of a multi-objective genetic algo...
详细信息
ISBN:
(纸本)9783642049613
multi-objective genetic algorithm is proved to be suitable for solving multi-objective optimization problems. However, it is usually very hard to balance the convergence and diversity of a multi-objective genetic algorithm. This paper introduces a new algorithm, with both good convergence and diversity based on clustering method and multi-parent crossover operator. Meanwhile, an initial population is generated by orthogonal design to enhance the search effort of the algorithm. The experimental results on a number of test problems indicate the good performance of the Cluster-Based Orthogonal multi-objective Genetic Algorithm.
An improved multi-objective genetic algorithm for structural passive control system optimization is proposed. Based on the two-branch tournament genetic algorithm, the selection operator is constructed by evaluating i...
详细信息
ISBN:
(纸本)9780819478047
An improved multi-objective genetic algorithm for structural passive control system optimization is proposed. Based on the two-branch tournament genetic algorithm, the selection operator is constructed by evaluating individuals according to their dominance in one run. For a constrained problem, the dominance-based penalty function method is advanced, containing information on an individual's status (feasible or infeasible), position in a search space, and distance from a Pareto optimal set. The proposed approach is used for the optimal designs of a six-storey building with shape memory alloy dampers subjected to earthquake. The number and position of dampers are chosen as the design variables. The number of dampers and peak relative inter-storey drift are considered as the objective functions. Numerical results generate a set of non-dominated solutions.
Support vector machines were originally proposed for the binary classification. For multiclass classification, some kinds of extensions of SVMs have been proposed. In this paper, we focus on "all together" m...
详细信息
ISBN:
(纸本)9784907764289
Support vector machines were originally proposed for the binary classification. For multiclass classification, some kinds of extensions of SVMs have been proposed. In this paper, we focus on "all together" method, where an extended SVM is constructed by using a piece-wise linear function. This model is formulated as an optimizationproblem which maximizes margins between each pair of classes for the generalization ability. However, as we point out in this paper, the model does not correctly represent the margins. Therefore, we propose a multi-objective model which exactly maximizes all margins. In addition, we derive a new SVM as a single-objective quadratic programming problem and apply the proposed SVM to some problems and verify its efficiency.
Innovative FRP-concrete composite bridge deck systems fabricated by casting concrete above FRP panel have been proposed and have already been implemented actually. In such deck systems, the FRP panel plays the role of...
详细信息
Innovative FRP-concrete composite bridge deck systems fabricated by casting concrete above FRP panel have been proposed and have already been implemented actually. In such deck systems, the FRP panel plays the role of formwork during the curing of concrete and behaves as a structural member combined with concrete during service life. Coarse sand coating is usually applied to secure composite behavior of the two materials instead of epoxy bonding, which is extensively used for retrofit purpose of deteriorated structures. This study focuses on investigating the bond behavior of coarse sand coating experimentally and analytically. Pure shear tests on coarse sand coated and epoxy coated joints are performed from which results are used to extract the bond-slip model. A method to obtain the bond-slip model from pure shear test is proposed, in which analytic solution of bond-slip model based on fracture mechanics is introduced and physical programming technique is used for formulation. Genetic algorithm is adopted for the optimization. Application of the proposed method produces local bond-slip model for coarse sand coating. Finally, in order to verify the proposed method and local bond-slip model, 4-point bending test of beam specimens with coarse sand coated FRP plate is performed and the result is compared to finite element analysis including the proposed local bond-slip model. (c) 2005 Elsevier Ltd. All rights reserved.
In this paper, a new multi-objective evolutionary algorithm for solving high complex multi-objectiveproblems is presented based on the rule of energy minimizing and the law of entropy increasing of particle systems i...
详细信息
ISBN:
(纸本)9781595931863
In this paper, a new multi-objective evolutionary algorithm for solving high complex multi-objectiveproblems is presented based on the rule of energy minimizing and the law of entropy increasing of particle systems in phase space, Through the experiments it proves that this algorithm can quickly obtains the Pareto solutions with high precision and uniform distribution. And the results of the experiments show that this algorithm can avoid the premature phenomenon of problems better than the traditional evolutionary algorithm because it can drive all the individuals to participate in the evolving operation in each generation.
In this study, an e-learning system is developed to handle the c-learning environment based on the learning ecological model. In the learning ecological model, which represents the comprehensive e-learning environment...
详细信息
In this study, an e-learning system is developed to handle the c-learning environment based on the learning ecological model. In the learning ecological model, which represents the comprehensive e-learning environment, not only the contents of learning, but also the learning environment are managed and provided, based on the content, the goal, and the configuration of the learning. The major purpose of this study is to realize the function that can manage the diversified learning objects with various information granularities and representation formats, using the learning object metadata, so that each learner can utilize the learning object based on the learning scenario, which is matched to the individual learner. The learning scenario is constructed by sequencing the learning objects based on the learning necessity, the learning history information, and the curriculum information of the object of learning, according to the characteristics of the learning object. As the sequencing procedure, the sequencing of the learning objects is considered, by applying the optimization technique of the multi-objective optimization problem, so that multiple evaluation viewpoints are simultaneously satisfied. The genetic algorithm is used as the optimization procedure. The learning object metadata and the sequencing of the learning objects are discussed in detail in this paper. The evaluation of the developed e-learning system is also described. (C) 2004 Wiley Periodicals, Inc.
暂无评论