Evolutionary algorithms (EA) are efficient population-based stochastic algorithms for solving optimizationproblems. The performance of EAs largely depends on the configuration of values of parameters that control the...
详细信息
ISBN:
(纸本)9781665487689
Evolutionary algorithms (EA) are efficient population-based stochastic algorithms for solving optimizationproblems. The performance of EAs largely depends on the configuration of values of parameters that control their search. Previous works studied how to configure EAs, though, there is a lack of a general approach to effectively tune EAs. To fill this gap, this paper presents a consistent, automated approach for tuning and controlling parameterized search of an EA. For this, we propose a deep reinforcement learning (DRL) based approach called `DRL-APC-DE' for online controlling search parameter values for a multi-objective Differential Evolution algorithm. The proposed method is trained and evaluated on widely adopted multi-objective test problems. The experimental results show that the proposed approach performs competitively to a non-adaptive Differential Evolution algorithm, tuned by grid search on the same range of possible parameter values. Subsequently, the trained algorithms have been applied to unseen multi-objectiveproblems for the adaptive control of parameters. Results show the successful ability of DRL-APC-DE to control parameters for solving these problems, which has the potential to significantly reduce the dependency on parameter tuning for the successful application of EAs.
To improve convergence performance of the algorithm and prevent the algorithm from falling into local optimal location, we proposes a novel fuzzy multi-objective particle swarm optimization based on linear differentia...
详细信息
ISBN:
(纸本)9783031096778;9783031096761
To improve convergence performance of the algorithm and prevent the algorithm from falling into local optimal location, we proposes a novel fuzzy multi-objective particle swarm optimization based on linear differential decline (LDDFMOPSO). In LDDFMOPSO, the fuzzy control strategy is applied to the inertia weight, so that the search ability of the global and local can be flexibly adjusted, thereby improving convergence performance of the algorithm. At the same time, in order to prevent the algorithm from falling into local optimal location, the strategy of linear differential decline is used to adjust the position change of particles. The experimental results illustrate that LDDFMOPSO has good performance compared to four state-of-the-art multi-objective particle swarm optimizations.
In real problems in Engineering, solving a problem is not enough;the solution of the problem must be the best solution possible. In other words, it is necessary to find the optimal solution. The solution is the best p...
详细信息
ISBN:
(纸本)9781509011476
In real problems in Engineering, solving a problem is not enough;the solution of the problem must be the best solution possible. In other words, it is necessary to find the optimal solution. The solution is the best possible solution because in the real world this problem may have certain constraints by which the solutions found may be feasible, that is, they can be implemented in practice and, unfeasible or that they cannot be implemented. Some of these problems in engineering can be MOP (multi-objectiveoptimization Problem). A general MOP includes a set of n parameters ( decision variables), a set of k objective functions and a set of m restrictions. The objective and restriction functions are functions of the decision variables where is possible to obtain a set of optimal values. Then the MOP can be expressed as: Optimize y = f(x) = (fl( x), f2(x),..., fk(x)) Subject to e(x) = (el(x), e2(x),..., em(x)) 0 Where x = (x1, x2,..., xn) X y = (y1, y2,..., yk) Y The method evolutionary algorithm (EA) refers to searching and optimization techniques based on the evolution model proposed by Charles Darwin. Genetic algorithms are used in several areas especially for searching and optimizations. In the real case the algorithm is implemented by choosing a coding for the possible solutions to the problem. The coding is done through chains of bits, numbers or characters that represent the chromosomes. The crossing and mutation operations are applied in a very simple way through functions of vector value manipulation. The EAs are interesting given the fact that at first glance they seem especially apt to deal with the difficulties presented by MOPs. The reason for this is that they can return an entire set of solutions after a simple run and they do not have any other of the limitations of traditional techniques. In addition, some researchers have suggested that the EAs would behave better than other blind searching techniques.
作者：
Zhu, ChenLiu, JingXidian Univ
Minist Educ Key Lab Intelligent Percept & Image Understanding Xian 710071 Peoples R China
A direction based multi-objective agent genetic algorithm (DMOAGA) is proposed in this paper. In order to take advantage of the effective direction information and depth of local search to mine non-dominated solutions...
详细信息
ISBN:
(纸本)9783642412783;9783642412776
A direction based multi-objective agent genetic algorithm (DMOAGA) is proposed in this paper. In order to take advantage of the effective direction information and depth of local search to mine non-dominated solutions, the direction perturbation operator is also employed. The neighborhood non-dominated solutions are generated using tournament selection and "average distance" rule, which maintains the diversity of non-dominated solution set. In the experiments, the benchmark problems UF1 similar to UF6 and ZDT1 similar to ZDT4 are used to validate the performance of DMOAGA. We compared it with NSGA-II and DMEA in terms of generational distance (GD) and inverted generational distance (IGD). The results show that DMOAGA has a good diversity and convergence, the performances on most of benchmark problems are better than DMEA and NSGA-II.
Particle swarm optimizer (PSO) is suitable for solving multi-objective optimization problems (MOPs). However, there are two main issues for any multi-objective particle swarm optimizers (MOPSOs). The first issue is ho...
详细信息
ISBN:
(纸本)9781665440899
Particle swarm optimizer (PSO) is suitable for solving multi-objective optimization problems (MOPs). However, there are two main issues for any multi-objective particle swarm optimizers (MOPSOs). The first issue is how to balance the convergence and diversity. The second issue is how to enhance the exploitation and exploration during the evolutionary procedure. In order to address these issues, an modified inverted generational distance (IGD(+)) performance indicator based PSO (IGD(+)-MOPSO) is proposed. The external archive updating strategies based on the IGD(+) indicator and the objective space decomposition method are proposed to select the evenly distributed non-dominated solutions. The leader updated strategy of each particle is based on the IGD(+) indicator value which is associated to the corresponding reference vector. The genetic operator is embedded into the evolutionary procedure to reset the position in order to help the particle jump out of the local optimum. We have conducted the simulation on some related benchmark test instances. The experimental results have indicated that the proposed algorithm is competitive with some related algorithms.
This paper proposes a novel voice adaptation method that we applied to interactive activities such as games where source and target data are unaligned. Conventional methods have seen the use of probabilistic models or...
详细信息
ISBN:
(纸本)9781728145334
This paper proposes a novel voice adaptation method that we applied to interactive activities such as games where source and target data are unaligned. Conventional methods have seen the use of probabilistic models or more recently, Deep Neural Networks. Common for most methods is that they require multiple subjects to train in conjunction, thus voice adaptation is not practical to be used in commercial applications. We propose a method which convert audible frequencies to light spectrum simple RGB color format, and not comparing sound signal similarities, but rather likeness in color. The comparison is done using multi-objectiveoptimization which considers raw and normalized frame colors as two separate objectives to be evaluated, respectively audible and spectral structure. The distance for the objectives is used to select an ideal output frame. Finally, prosodic information such as speech intensity is translated from measured input values onto the designated output frame. The method is evaluated using MOS, ABX, performance benchmark and lastly implemented into the Unity3D game engine as a proof of concept. Results show good sound quality and high performance with little output fragmentation.
Dynamic multi -objectiveoptimizationproblems (DMOPs) are multi -objectiveoptimizationproblems in which at least one objective and/or related parameter vary over time. The challenge of solving DMOPs is to efficient...
详细信息
Dynamic multi -objectiveoptimizationproblems (DMOPs) are multi -objectiveoptimizationproblems in which at least one objective and/or related parameter vary over time. The challenge of solving DMOPs is to efficiently and accurately track the true Pareto-optimal set when the environment undergoes changes. However, many existing prediction -based methods overlook the distinct individual movement directions and the available information in the objective space, leading to biased predictions and misleading the subsequent search process. To address this issue, this paper proposes a prediction method called IMDMOEA, which relies on cluster center points and induced mutation. Specifically, employing linear prediction methods based on cluster center points in the decision space enables the algorithm to rapidly capture the population's evolutionary direction and distributional shape. Additionally, to enhance the algorithm's adaptability to significant environmental changes, the induced mutation strategy corrects the population's evolutionary direction by selecting promising individuals for mutation based on the predicted result of the Pareto front in the objective space. These two complementary strategies enable the algorithm to respond faster and more effectively to environmental changes. Finally, the proposed algorithm is evaluated using the JY, dMOP, FDA, and F test suites. The experimental results demonstrate that IMDMOEA competes favorably with other state-of-the-art algorithms.
As a powerful optimization technique, multi-objective particle swarm optimization (MOPSO) has been paid more and more attention by scientists. However, in more complex problems, MOPSO faces the challenges of weak glob...
详细信息
As a powerful optimization technique, multi-objective particle swarm optimization (MOPSO) has been paid more and more attention by scientists. However, in more complex problems, MOPSO faces the challenges of weak global search ability and easy-to-fall-into local optimality. To address these challenges and obtain better solutions, people have proposed many variants. In this study, a density-guided and adaptive update strategy for multi-objective particle swarm optimization (DAMOPSO) is proposed. First, an adaptive grid is used to determine the mutation particles and guides. Then, the Cauchy mutation operator is performed for the poorly distributed particles to expand the search space of the population. Additionally, the strategy of non-dominated sorting and hyper-region density are devised for maintaining external archives, which contribute to the uniform distribution of optimal solutions. Finally, an adaptive detection strategy based on the adjustment coefficient and conversion efficiency is designed to update the flight parameters. These approaches not only speed up the convergence of algorithms, but also balance exploitation and exploration more effectively. The proposed algorithm is compared with several representative multi-objectiveoptimization algorithms on 22 benchmark functions;meanwhile, statistical tests, ablation experiments, analysis of stability, and complexity are also performed. The experimental results demonstrate DAMOPSO is more competitive than other comparison algorithms. Graphical Abstract
This paper describes a new approach to register-pressure-aware instruction scheduling, using Ant Colony optimization (ACO). ACO is a nature-inspired optimization technique that researchers have successfully applied to...
详细信息
This paper describes a new approach to register-pressure-aware instruction scheduling, using Ant Colony optimization (ACO). ACO is a nature-inspired optimization technique that researchers have successfully applied to NP-hard sequencing problems like the Traveling Salesman Problem (TSP) and its derivatives. In this work, we describe an ACO algorithm for solving the long-standing compiler optimization problem of balancing Instruction-Level Parallelism (ILP) and Register Pressure (RP) in pre-allocation instruction scheduling. Three different cost functions are studied for estimating RP during instruction scheduling. The proposed ACO algorithm is implemented in the LLVM open-source compiler, and its performance is evaluated experimentally on three different machines with three different instruction-set architectures: Intel x86, ARM, and AMD GPU. The proposed ACO algorithm is compared to an exact Branch-and-Bound (B&B) algorithm proposed in previous work. On x86 and ARM, both algorithms are evaluated relative to LLVM's generic scheduler, while on the AMD GPU, the algorithms are evaluated relative to AMD's production scheduler. The experimental results show that using SPECrate 2017 Floating Point, the proposed algorithm gives geometric-mean improvements of 1.13% and 1.25% in execution speed on x86 and ARM, respectively, relative to the LLVM scheduler. Using PlaidML on an AMD GPU, it gives a geometric-mean improvement of 7.14% in execution speed relative to the AMD scheduler. The proposed ACO algorithm gives approximately the same execution-time results as the B&B algorithm, with each algorithm outperforming the other on a substantial number of hard scheduling regions. ACO gives better results than B&B on many large instances that B&B times out on. Both ACO and B&B outperform the LLVM algorithm on the CPU and the AMD algorithm on the GPU.
In this article, Artificial Cooperative Search (ACS) algorithm is incorporated with the quadratic approximation (QA) operator to solve the multi-objective economic emission load dispatch (EELD) problems with different...
详细信息
In this article, Artificial Cooperative Search (ACS) algorithm is incorporated with the quadratic approximation (QA) operator to solve the multi-objective economic emission load dispatch (EELD) problems with different generation units. ACS is a Swarm Intelligence-based metaheuristic algorithm, based on the interaction between prey and predator organisms in a habitat, which is effective at global search;however, it does not perform so well at exploring promising regions. The QA operator, on the other hand, is a non-derivative-based efficient local search method that finds the minimum of a quadratic hyperspace passing through three points in a D-dimensional space. Solving the EELD problems with the hybridized ACS-QA algorithm, as being proposed in the present article, leads to more accurate results with fewer function evaluations. Also, multi-objectivity of the problem is handled by transforming it into a single-objective problem by using the weighted sum method. The efficiency of the proposed ACS-QA algorithm is tested in comparison to the algorithms existing in literature by implementing it on six different benchmark optimizationproblems. Afterward, the proposed ACS-QA algorithm and the ACS algorithm are implemented on multi-objective EELD problems with different generation units. The results are compared with the solutions in literature utilizing different metaheuristic optimization algorithms. Both studies firmly showed that the ACS-QA algorithm is able to find more accurate results even though it uses fewer function evaluation calls.
暂无评论