scholarly journals Multi-Objective Evolutionary Instance Selection for Regression Tasks

Entropy ◽  
2018 ◽  
Vol 20 (10) ◽  
pp. 746 ◽  
Author(s):  
Mirosław Kordos ◽  
Krystian Łapa

The purpose of instance selection is to reduce the data size while preserving as much useful information stored in the data as possible and detecting and removing the erroneous and redundant information. In this work, we analyze instance selection in regression tasks and apply the NSGA-II multi-objective evolutionary algorithm to direct the search for the optimal subset of the training dataset and the k-NN algorithm for evaluating the solutions during the selection process. A key advantage of the method is obtaining a pool of solutions situated on the Pareto front, where each of them is the best for certain RMSE-compression balance. We discuss different parameters of the process and their influence on the results and put special efforts to reducing the computational complexity of our approach. The experimental evaluation proves that the proposed method achieves good performance in terms of minimization of prediction error and minimization of dataset size.

Author(s):  
Rui Zhu ◽  
Puyu Cao ◽  
Yang Wang ◽  
Chao Ning

Abstract Flow distortions occur at the outlet section of the intake duct owing to its shape properties, which is a component of water-jet propulsion. Since the noticeable influence of intake’s flow characteristics upon propulsive efficiency, it’s necessary to focus on intake duct redesign. In this paper, a systematic methodology for reducing flow distortions and power losses within the intake duct through a shape optimization process was obtained. In addition, the mechanism of flow distortions was also developed. The flush type inlet applied in the marine vessel with the speed of 30 knots was chosen as research project. Four characteristic parameters were set as optimization variables depending on the geometrical relationship of thirteen characteristic parameters referred to the duct longitudinal midsection, which were the ramp angle α, the radius of the upper lip R3, the radius of the lower lip R4 and the lip height h respectively. Subsequently, a sample space was built by Latin Hypercube Sampling (LHS) and the parameters were normalized in the range of 0 to 1. With the commercial software CFX, the numerical simulation was accomplished driven by SST k-ω turbulence model. Multi-objective optimization based on the Non-Dominated Sorting Genetic Algorithm II (NSGA-II) was utilized to minimize the non-uniformity at outlet section and maximize the minimal pressure at lip simultaneously. Moreover, the Radial Basis Function (RBF) neural network was employed to approximate the functional relationship between variables and objectives, which could be applied in the NSGA-II to get the Pareto Front. The minimum non-uniformity point and the trade-off point (The point both satisfies the minimum non-uniformity and the maximum minimal pressure at lip strategically) were selected from the Pareto Front. With regard to the characteristic parameters of the trade-off point, the ramp angle, the radius of the upper lip, the radius of the lower lip and the lip height are 31.91°, 11.42 mm, 400.97 mm and 55.43 mm respectively. Meanwhile, the characteristic parameters of the minimum non-uniformity point are 30.22°, 25.59 mm, 166.65 mm and 89.90 mm respectively. Ultimately, the duct outflow characteristics of prototype and optimization are compared. In terms of the trade-off point, the minimal pressure at lip increases 66.40% to −24488.93 Pa and the non-uniformity has a drop of 4.56% to 0.1571. The non-uniformity of the minimum point is 0.1481 which is reduced by 10.02%. Through the optimization of duct shape, the secondary flow (Dean vortices) is suppressed effectively. This paper is expected to provide a better comprehension of the flow field within the intake duct of water-jet propulsion.


2014 ◽  
Vol 945-949 ◽  
pp. 2241-2247
Author(s):  
De Gao Zhao ◽  
Qiang Li

This paper deals with application of Non-dominated Sorting Genetic Algorithm with elitism (NSGA-II) to solve multi-objective optimization problems of designing a vehicle-borne radar antenna pedestal. Five technical improvements are proposed due to the disadvantages of NSGA-II. They are as follow: (1) presenting a new method to calculate the fitness of individuals in population; (2) renewing the definition of crowding distance; (3) introducing a threshold for choosing elitist; (4) reducing some redundant sorting process; (5) developing a self-adaptive arithmetic cross and mutation probability. The modified algorithm can lead to better population diversity than the original NSGA-II. Simulation results prove rationality and validity of the modified NSGA-II. A uniformly distributed Pareto front can be obtained by using the modified NSGA-II. Finally, a multi-objective problem of designing a vehicle-borne radar antenna pedestal is settled with the modified algorithm.


2016 ◽  
Vol 23 (5) ◽  
pp. 782-793 ◽  
Author(s):  
Mansour Ataei ◽  
Ehsan Asadi ◽  
Avesta Goodarzi ◽  
Amir Khajepour ◽  
Mir Behrad Khamesee

This paper reports work on the optimization and performance evaluation of a hybrid electromagnetic suspension system equipped with a hybrid electromagnetic damper. The hybrid damper is configured to operate with hydraulic and electromagnetic components. The hydraulic component produces a large fail-safe baseline damping force, while the electromagnetic component adds energy regeneration and adaptability to the suspension. For analyzing the system, the electromagnetic component was modeled and integrated into a 2DOF quarter-car model. Three criteria were considered for evaluating the performance of the suspension system: ride comfort, road holding and regenerated power. Using the genetic algorithm multi-objective optimization (NSGA-II), the suspension design was optimized to improve the performance of the vehicle with respect to the selected criteria. The multi-objective optimization method provided a set of solutions called Pareto front in which all solutions are equally good and the selection of each one depends on conditions and needs. Among the given solutions in the Pareto front, a small number of cases, with different design purposes, were selected. The performances of the selected designs were compared with two reference systems: a conventional and a nonoptimized hybrid suspension system. The results show that the ride comfort and road holding qualities of the optimized hybrid system are improved, and the regenerated power is considerably increased.


Author(s):  
Minami Miyakawa ◽  
◽  
Keiki Takadama ◽  
Hiroyuki Sato

As an evolutionary approach to solve multi-objective optimization problems involving several constraints, recently a multi-objective evolutionary algorithm (MOEA) using two-stage non-dominated sorting and directed mating (TNSDM) has been proposed. In TNSDM, directed mating utilizes infeasible solutions dominating feasible solutions in the objective space to generate offspring. In our previous studies, significant contribution of directed mating to the improvement of the search performancewas verified on several benchmark problems. However, in the conventional TNSDM, infeasible solutions utilized in directed mating are discarded in the selection process of parents (elites) population and cannot be utilized in the next generation. TNSDM has potential to further improve the search performance by archiving useful solutions for directed mating to the next generation and repeatedly utilizing them in directed mating. To further improve effects of directed mating in TNSDM, in this work, we propose an archiving strategy of useful solutions for directed mating. We verify the search performance of TNSDM using the proposed archive by varying the size of archive, and compare its search performance with the conventional CNSGA-II and RTS onmobjectiveskknapsacks problems. As results, we show that the search performance of TNSDM is improved by introducing the proposed archive in aspects of diversity of the obtained solutions in the objective space and convergence of solutions toward the optimal Pareto front.


Author(s):  
Zhongran Chi ◽  
Haiqing Liu ◽  
Shusheng Zang

This paper discusses the approach of cooling design optimization of a high-pressure turbine (HPT) endwall with applied 3D conjugate heat transfer (CHT) computational fluid dynamics (CFD). This study involved the optimization of the spacing of impingement jet array and the exit width of shaped holes, which are different for each cooling cavity. The optimization objectives were to reduce the wall-temperature level and to increase the aerodynamic performance. The optimization methodology consisted of an in-house parametric design and CFD mesh generation tool, a CHT CFD solver, a database of CFD results, a metamodel, and an algorithm for multi-objective optimization. The CFD tool was validated against experimental data of an endwall at CHT conditions. The metamodel, which could efficiently estimate the optimization objectives of new individuals without CFD runs, was developed and coupled with nondominated sorting genetic algorithm II (NSGA II) to accelerate the optimization process. Through the optimization search, the Pareto front of the problem was found in each iteration. The accuracy of metamodel with more iterations was improved by enriching database. But optimal designs found by the last iteration are almost identical with those of the first iteration. Through analyzing extra CFD results, it was demonstrated that the design variables in the Pareto front successfully reached the optimal values. The optimal pitches of impingement arrays could be decided accommodating the local thermal load while avoiding jet lift-off of film coolant. It was also suggested that cylindrical film holes near throat should be beneficial to both aerodynamic and cooling performances.


2020 ◽  
Vol 25 (2) ◽  
pp. 32 ◽  
Author(s):  
Gustavo-Adolfo Vargas-Hákim ◽  
Efrén Mezura-Montes ◽  
Edgar Galván

This work presents the assessment of the well-known Non-Dominated Sorting Genetic Algorithm II (NSGA-II) and one of its variants to optimize a proposed electric power production system. Such variant implements a chaotic model to generate the initial population, aiming to get a better distributed Pareto front. The considered power system is composed of solar, wind and natural gas power sources, being the first two renewable energies. Three conflicting objectives are considered in the problem: (1) power production, (2) production costs and (3) CO2 emissions. The Multi-Objective Evolutionary Algorithm based on Decomposition (MOEA/D) is also adopted in the comparison so as to enrich the empirical evidence by contrasting the NSGA-II versions against a non-Pareto-based approach. Spacing and Hypervolume are the chosen metrics to compare the performance of the algorithms under study. The obtained results suggest that there is no significant improvement by using the variant of the NSGA-II over the original version. Nonetheless, meaningful performance differences have been found between MOEA/D and the other two algorithms.


2012 ◽  
Vol 2012 ◽  
pp. 1-11 ◽  
Author(s):  
Yufang Qin ◽  
Junzhong Ji ◽  
Chunnian Liu

Multiobjective optimization problem (MOP) is an important and challenging topic in the fields of industrial design and scientific research. Multi-objective evolutionary algorithm (MOEA) has proved to be one of the most efficient algorithms solving the multi-objective optimization. In this paper, we propose an entropy-based multi-objective evolutionary algorithm with an enhanced elite mechanism (E-MOEA), which improves the convergence and diversity of solution set in MOPs effectively. In this algorithm, an enhanced elite mechanism is applied to guide the direction of the evolution of the population. Specifically, it accelerates the population to approach the true Pareto front at the early stage of the evolution process. A strategy based on entropy is used to maintain the diversity of population when the population is near to the Pareto front. The proposed algorithm is executed on widely used test problems, and the simulated results show that the algorithm has better or comparative performances in convergence and diversity of solutions compared with two state-of-the-art evolutionary algorithms: NSGA-II, SPEA2 and the MOSADE.


Author(s):  
Zhongran Chi ◽  
Haiqing Liu ◽  
Shusheng Zang

This paper discusses the approach of cooling design optimization of a HPT endwall with 3D Conjugate Heat Transfer (CHT) CFD applied. This study involved the optimization of the spacing of impingement jet array and the exit width of shaped holes, which were different for each cooling cavity. The optimization objectives were to reduce the wall temperature level and also to increase the aerodynamic performance of the gas turbine. The optimization methodology consisted of an in-house parametric design & CFD mesh generation tool, a CHT CFD solver, a database of wall temperature distributions, a metamodel, and a genetic algorithm (GA) for evolutionary multi-objective optimization. The CFD tool was validated against experimental data of an endwall at CHT conditions. The metamodel, which could efficiently predict the aerodynamic loss and the wall temperature distribution of a new individual based on the database, was developed and coupled with Non-dominated Sorting Genetic Algorithm II (NSGA-II) to accelerate the optimization process. Through optimization search, the Pareto front of the problem was found costing only tens of CFD runs. By comparing with additional CFD results, it was demonstrated that the design variables in the Pareto front successfully reached the optimal values. The optimal spacing of each impingement array was decided accommodating the local thermal load while avoiding jet lift-off of film coolant. It was also suggested that using cylindrical film holes near throat could benefit both aerodynamics and cooling.


2019 ◽  
Vol 19 (8) ◽  
pp. 2338-2346
Author(s):  
Silvia Carpitella ◽  
Bruno Brentan ◽  
Idel Montalvo ◽  
Joaquín Izquierdo ◽  
Antonella Certa

Abstract This work presents a multi-criteria-based approach to automatically select specific non-dominated solutions from a Pareto front previously obtained using multi-objective optimization to find optimal solutions for pump control in a water supply system. Optimal operation of pumps in these utilities is paramount to enable water companies to achieve energy efficiency in their systems. The Fuzzy Technique for Order of Preference by Similarity to Ideal Solution (FTOPSIS) is used to rank the Pareto solutions found by the non-dominated sorting genetic algorithm (NSGA-II) employed to solve the multi-objective problem. Various scenarios are evaluated under leakage uncertainty conditions, resulting in fuzzy solutions for the Pareto front. This paper shows the suitability of the approach for quasi real-world problems. In our case-study, the obtained solutions for scenarios including leakage represent the best trade-off among the optimal solutions, under some considered criteria, namely, operational cost, operational lack of service, pressure uniformity and network resilience. Potential future developments could include the use of clustering alternatives to evaluate the goodness of each solution under the considered evaluation criteria.


Water ◽  
2021 ◽  
Vol 13 (10) ◽  
pp. 1421
Author(s):  
Jisi Fu ◽  
Ping-An Zhong ◽  
Bin Xu ◽  
Feilin Zhu ◽  
Juan Chen ◽  
...  

Transboundary water resources allocation is an effective measure to resolve water-related conflicts. Aiming at the problem of water conflicts, we constructed water resources allocation models based on game theory and multi-objective optimization, and revealed the differences between the two models. We compare the Pareto front solved by the AR-MOEA method and the NSGA-II method, and analyzed the difference between the Nash–Harsanyi Leader–Follower game model and the multi-objective optimization model. The Huaihe River basin was selected as a case study. The results show that: (1) The AR-MOEA method is better than the NSGA-II method in terms of the diversity metric (Δ); (2) the solution of the asymmetric Nash–Harsanyi Leader–Follower game model is a non-dominated solution, and the asymmetric game model can obtain the same water resources allocation scheme of the multi-objective optimal allocation model under a specific preference structure; (3) after the multi-objective optimization model obtains the Pareto front, it still needs to construct the preference information of the Pareto front for a second time to make the optimal solution of a multi-objective decision, while the game model can directly obtain the water resources allocation scheme at one time by participating in the negotiation. The results expand the solution method of water resources allocation models and provide support for rational water resources allocation.


Sign in / Sign up

Export Citation Format

Share Document