M.C. Garrido

Using machine learning in a cooperative hybrid parallel strategy of metaheuristics

This paper proposes the construction of a centralized hybrid metaheuristic cooperative strategy to solve optimization problems. Knowledge (intelligence) is incorporated into the coordinator to improve performance. This knowledge is incorporated through a set of rules and models obtained from a knowledge extraction process applied to the records of the results returned by individual metaheuristics. The effectiveness of the approach is tested in several computational experiments in which we compare the results obtained by the individual metaheuristics, by several non-cooperative and cooperative strategies and by the strategy proposed in this paper.

Fundamentals for Design and Construction of a Fuzzy Random Forest

Following Breiman’s methodology, we propose the fundamentals to design and construct a “forest” of randomly generated fuzzy decision trees, i.e., a Fuzzy Random Forest. This approach combines the robustness of multi-classifiers, the construction efficiency of decision trees, the power of the randomness to increase the diversity of the trees in the forest, and the flexibility of fuzzy logic and the fuzzy sets for data managing. A prototype for the method has been constructed and we have implemented some specific strategies for inference in the Fuzzy Random Forest. Some experimental results are given.

Facing dynamic optimization using a cooperative metaheuristic configured via fuzzy logic and SVMs

Abstract: Many dynamic optimization problems appear in the real world, and to solve them we need to find strategies that can track the optimum as it moves in the search space. In this paper we propose the use of a cooperative metaheuristic to cope with such problems. In this strategy different metaheuristics cooperate under the supervision of a coordinator. This coordinator is able to control the cooperation using a collection of Support Vector Machine models and a fuzzy decision framework. The combination of these two techniques allows us to modify the behavior of the strategy depending on the instance being solved. In order to obtain the models we use a well defined knowledge extraction process, which is performed only once and before operating the strategy. To test the validity of this approach we have applied it to a combinatorial optimization problem and a continuous optimization problem, respectively, Dynamic Knapsack Problem and Moving Peaks Benchmark. The strategy has been compared with other individual and cooperative strategies with very interesting results.

Combination methods in a Fuzzy Random Forest

When individual classifiers are combined appropriately, we usually obtain a better performance in terms of classification precision. Multi-classifiers are the result of combining several individual classifiers. In this work we propose and compare various combination methods to obtain the final decision of the multi-classifier based on a ldquoforestrdquo of randomly generated fuzzy decision trees, i.e., a Fuzzy Random Forest. We propose various forms of weighting decisions on the basis of information obtained from the FRF. We make a comparative study with several databases to show the efficiency of the various combination methods.

A Hybrid System of Nature Inspired Metaheuristics

Nature-inspired metaheuristics are effective strategies for solving optimization problems. However, when trying to solve an instance of this kind of problems it is hard to know which algorithm should be used (algorithm-instance problem). Hybrid systems provide flexible tools that can help to cope with this problem. Therefore a hybrid system based on the intelligent combination of different natureinspired strategies will give more robustness and will allow to find higher quality solutions for different instance types.

A Cooperative System of Metaheuristics

Hybrid systems give more flexible mechanisms for solving complex problems that can be very difficult to solve using less tolerant approaches. Therefore, a hybrid system will be the most suitable tool in order to cope with the algorithm-instance problem, which says that it is possible that an algorithm and its parameters that obtain good results for an instance of a problem, do not get the same results for another instance of the same problem. All this leads us to use different algorithms to solve combinatorial optimization problems within a single coordinated schema, that is a hybrid cooperative system of metaheuristics. In order to build this system we have proposed a methodology for the construction of a hybrid system, based on data mining and soft computing. In order to test the usefulness of this methodology two hybrid systems based on a fuzzy model have been constructed to solve the knapsack problem. The first system coordinates two metaheuristics, a genetic algorithm and a tabu search. The second one adds a third metaheuristic, simulated annealing, in order to check the robustness of the system and its capacity of obtaining higher quality solutions when a metaheuristic is added. Results obtained by this systems and a comparison with the ones obtained with individual metaheuristics are shown.

Fuzzy integral in systems modeling

When we wish to model a system from a set of examples we use a method that is capable of supporting and modeling those examples. There are many candidates and the question may always arise as to which is performing best or whether a combination is providing better results. In this paper we present a method for using fuzzy integral as a fusion operator both for the different types of modeling methods, and as a selection operator capable of informing as to the importance of the sources, which are used. Thus, we are left with a reduced set, which improves the modeling of the system as far as is possible.

Improving fuzzy pattern matching techniques to deal with non discrimination ability features

Fuzzy pattern matching technique represents a group of fuzzy methods for supervised fuzzy pattern recognition. It has a number of advantages over other pattern recognition methods, including simpler methods of feature selection or ability to learn in real time environments, but its main drawback is it is not able to model the correlation between features, since fuzzy pattern matching assumes non interactivity between them. This paper presents an attempt to extend this technique to deal with this kind of features. To show the accuracy of the proposed solution, we present the results obtained in a simulated data set (an extension of the xor problem) and a real data set (the Wisconsin breast cancer data set).

Impact of Fuzzy Logic in the Cooperation of Metaheuristics

Algorithm selection problem is a common problem when we solve optimization problems. To cope with it we have proposed a hybrid system of metaheuristics that intelligently combines different strategies using a coordinator based on Fuzzy Logic. In this paper we study the impact of Fuzzy Logic in the behaviour of this hybrid system. In order to do that we perform some test to study the impact of an important parameter, the α− cut used in the fuzzy engine of the system, demonstrating how the variations on this parameter may change the performance of the system with different kind of instances.

Efficiency of the mixture model components using fuzzy integrals

In this paper -we describe the fuzzy integral used not only as a fusion operator, but also as a selection/removal/add operator. This operator is able to tell us about the importance of the sources that we merge. In the system modeling context, we have applied this technique to mixture models, where the number of component densities of the mixture model is not known a priori. Given a mixture model, we use the fuzzy integral like selection/removal/add operator of the component densities of this model. Hence, we will ascertain the number of components necessary with a controlled error. In order to obtain these, results, we propose a decision process based on coefficients provided by the fuzzy integral. We have used several data sets to evaluate the accuracy of the method. We show an illustrative example