Extracting comprehensible and general classifiers from data in the form of rule systems is an important task in many problem domains. This study investigates the utility of a multi-objectiveevolutionary algorithm (MO...
详细信息
Extracting comprehensible and general classifiers from data in the form of rule systems is an important task in many problem domains. This study investigates the utility of a multi-objectiveevolutionary algorithm (MOEA) for this task. multi-objective evolutionary algorithms are capable of finding several trade-off solutions between different objectives in a single run. In the context of the present study, the objectives to be optimised are the complexity of the rule systems, and their fit to the data. Complex rule systems are required to fit the data well. However, overly complex rule systems often generalise poorly on new data. In addition they tend to be incomprehensible. It is, therefore, important to obtain trade-off solutions that achieve the best possible fit to the data with the lowest possible complexity. The rule systems produced by the proposed multi-objectiveevolutionary algorithm are compared with those produced by several other existing approaches for a number of benchmark datasets. It is shown that the algorithm produces less complex classifiers that perform well on unseen data. (c) 2005 Elsevier Ireland Ltd. All rights reserved.
In this contribution we propose a multi-objectiveevolutionary algorithm for Tuning Fuzzy Rule-Based Systems by considering two objectives, accuracy and interpretability. To this aim we define a new objective that all...
详细信息
ISBN:
(纸本)9789899507968
In this contribution we propose a multi-objectiveevolutionary algorithm for Tuning Fuzzy Rule-Based Systems by considering two objectives, accuracy and interpretability. To this aim we define a new objective that allows preserving the interpretability of the system. This new objective is an interpretability index which is the union of three metrics to preserve the original shapes of the membership functions as much as possible while a tuning of the membership function parameters is performed. The proposed method has been compared to a single objective accuracy-guided algorithm in two real problems showing that many solutions in the Pareto front dominate to those obtained by the single objective-based one.
The calculation of ballast plan in load-out operations is generally performed assuming a rigid barge. This assumption may not be reliable since in reality the barge is flexible. Having incorrect ballast plan may lead ...
详细信息
The calculation of ballast plan in load-out operations is generally performed assuming a rigid barge. This assumption may not be reliable since in reality the barge is flexible. Having incorrect ballast plan may lead to overstressing of the loaded-out structure. We present a method of finding a more accurate ballast plan, taking into account the flexibility of the barge. This method makes use of a multi-objectiveevolutionary algorithm to find the optimum ballast arrangement at every load-out stage. We model the load-out configuration as a beam on elastic foundation, loaded with distributed trapezoidal loading representing the load from the structure. Minimizing deflection and curvature of the beam, as well as maximizing the ballast transfer efficiency between the load-out stages are chosen as the objectives of the algorithm. It is shown that the proposed method is better than the conventional rigid barge method in terms of minimizing the deflection and curvature as well as maximizing the ballast transfer efficiency.
In the last years, several papers have proposed to adopt multi-objective evolutionary algorithms (MOEAs) to generate Mamdani fuzzy rule-based systems with different trade-offs between interpretability and accuracy. Si...
详细信息
ISBN:
(纸本)9783642023187
In the last years, several papers have proposed to adopt multi-objective evolutionary algorithms (MOEAs) to generate Mamdani fuzzy rule-based systems with different trade-offs between interpretability and accuracy. Since interpretability is difficult to quantify because of its qualitative nature, several measures have been introduced, but there is no general agreement on any of them. In this paper, we propose all MOEA to learn concurrently rule base and membership function parameters by optimizing accuracy and interpretability, which is measured in terms of number of conditions in the antecedents of rules and partition integrity. Partition integrity is evaluated by using a purposely-defined index based on the piecewise linear transformation exploited to learn membership function parameters. Results on a real-world regression problem are shown and discussed.
This paper presents, for the first time, the application of the strength Pareto evolutionary algorithm to the multi-objective design of isolated hybrid systems, minimising both the total cost throughout the useful lif...
详细信息
This paper presents, for the first time, the application of the strength Pareto evolutionary algorithm to the multi-objective design of isolated hybrid systems, minimising both the total cost throughout the useful life of the installation and the unmet load. For this task. a multi-objectiveevolutionary algorithm (MOEA) and a genetic algorithm (GA) have been used in order to find the best combinations of components for the hybrid system and control strategy. Also, a novel control strategy has been developed and it will be expounded in this article. As an example of application, a PV-wind-diesel system has been designed, obtaining a Set of possible Solutions (Pareto set) from which the designer can choose those which he/she prefers considering the costs and unmet load of each. The results obtained demonstrate the practical utility of the design method used. (C) 2008 Elsevier B.V. All rights reserved.
The performance of the Dynamic Weight Aggregation system as applied to a Genetic Algorithm (DWAGA) and NSGA-II are evaluated and compared against each other. The algorithms are run on 11 test functions. The performanc...
详细信息
ISBN:
(纸本)0889864799
The performance of the Dynamic Weight Aggregation system as applied to a Genetic Algorithm (DWAGA) and NSGA-II are evaluated and compared against each other. The algorithms are run on 11 test functions. The performance of the algorithms is evaluated by examining the spacing, diversity and coverage of the Pareto front, as well as each algorithm's execution time. It is discovered that, while the NSGA-II performs better on most of the test functions, the DWAGA can outperform the NSGA-II on some of the functions, including a concave one.
A quality-time analysis of multi-objective evolutionary algorithms (MOEAs) based on schema theorem and building blocks hypothesis is developed. A bicriteria OneMax problem, a hypothesis of niche and species, and a def...
详细信息
ISBN:
(纸本)1595930108
A quality-time analysis of multi-objective evolutionary algorithms (MOEAs) based on schema theorem and building blocks hypothesis is developed. A bicriteria OneMax problem, a hypothesis of niche and species, and a definition of dissimilar schemata are introduced for the analysis. In this paper, the convergence time, the first and last hitting time models are constructed for analyzing the performance of MOEAs. Population sizing model is constructed for determining appropriate population sizes. The models are verified using the bicriteria OneMax problem. The theoretical results indicate how the convergence time and population size of a MOEA scale up with the problem size, the dissimilarity of Pareto-optimal solutions, and the number of Pareto-optimal solutions of a multi-objective optimization problem.
This paper presents a novel approach to support the selection of conceptual solutions to multi-objective problems. The proposed method involves a comparison between concepts, based on the performances of sets of solut...
详细信息
This paper presents a novel approach to support the selection of conceptual solutions to multi-objective problems. The proposed method involves a comparison between concepts, based on the performances of sets of solutions that represent them. The set-based comparison of concepts is consistent with the so-called Toyota set-based concurrent engineering process. Such an approach discourages early exploitation of solutions and promotes extended exploration of the design space by means of sets of solutions. Both optimality and variability of concepts are considered, and their measures are devised to pose the selection problem as an auxiliary multi-objective problem. The auxiliary objectives are to maximise optimality and to maximise the variability. This highlights the inherent multi-objectivity of concept selection and supports decision-making under the possible contradictory nature of optimality and variability of concepts. Both academic and engineering problems are used to demonstrate the approach and to expose the inherent subjectivity of the measures, which are dependent on the selection of a window of interest by the decision-makers.
The selection of descriptor subsets for QSAR/QSPR is a hard combinatorial problem that requires the evaluation of complex relationships in order to assess the relevance of the selected subsets. In this paper, we descr...
详细信息
The selection of descriptor subsets for QSAR/QSPR is a hard combinatorial problem that requires the evaluation of complex relationships in order to assess the relevance of the selected subsets. In this paper, we describe the main issues in applying descriptor selection for QSAR methods and propose a novel two-phase methodology for this task. The first phase makes use of a multi-objectiveevolutionary technique which yields interesting advantages compared to mono-objective methods. The second phase complements the first one and it enables to refine and improve the confidence in the chosen subsets of descriptors. This methodology allows the selection of subsets when a large number of descriptors are involved and it is also Suitable for linear and nonlinear QSAR/QSPR models. The proposed method was tested using three data sets with experimental values for blood-brain barrier penetration, human intestinal absorption and hydrophobicity. Results reveal the capability of the method for achieving subsets of descriptors with a high predictive capacity and a low cardinality. Therefore, our proposal constitutes a new promising technique helpful for the development of QSAR/QSPR models.
Formulating search queries based on a thematic context is a challenging problem due to the large number of combinations in which terms can be used to reflect the topic of interest. This paper presents a novel approach...
详细信息
ISBN:
(纸本)9783642049569
Formulating search queries based on a thematic context is a challenging problem due to the large number of combinations in which terms can be used to reflect the topic of interest. This paper presents a novel approach to learn topical queries that simultaneously satisfy multiple retrieval objectives. The proposed method consists in using a topic ontology to train an evolutionary Algorithm that incrementally moves a population of queries towards the proposed objectives. We present an analysis of different single- and multi-objective strategies, discuss their strengths and limitations and test the most promising strategies on a large set of labeled Web pages. Our evaluations indicate that the tested strategies that apply multi-objective evolutionary algorithms are significantly superior to a baseline approach that attempts to generate queries directly from a topic description.
暂无评论