Carlos Cruz Corona

Efficient multi-swarm PSO algorithms for dynamic environments

Particle swarm optimization has been successfully applied in many research and application areas because of its effectiveness and easy implementation. In this work we extend one of its variants to address multi-modal dynamic optimization problems, the multi-swarm PSO (mPSO) proposed by Blackwell and Branke. The aim of our proposal is to increase the efficiency of this algorithm. To this end, we propose techniques operating at swarm level: one of which divides each swarm into two groups depending on the quality of the particles for facing the loss of diversity, and the other control the number of active swarms during the run using a fuzzy rule. A detailed experimental analysis shows the robustness of our proposal.

Self-adaptive, multipopulation differential evolution in dynamic environments

The present work proposes a simple but effective self-adaptive strategy to control the behaviour of a differential evolution (DE) based multipopulation algorithm for dynamic environments. Specifically, the proposed scheme is aimed to control the creation of random individuals by the self-adaptation of the involved parameter. An interaction scheme between random and conventional DE individuals is also proposed and analyzed. The conducted computational experiments show that self-adaptation is profitable, leading to an algorithm that is as competitive as other efficient methods and able to beat the winner of the CEC 2009 competition on dynamic environments.

Improvement Strategies for Multi-swarm PSO in Dynamic Environments

Many real world optimization problems are dynamic, meaning that their optimal solutions are time-varying. In recent years, an effective approach to address these problems has been the multi-swarmPSO (mPSO). Despite this, we believe that there is still room for improvement and, in this contribution we propose two simple strategies to increase the effectiveness of mPSO. The first one faces the diversity loss in the swarm after an environment change; while the second one increases the efficiency through stopping swarms showing a bad behavior. From the experiments performed on the Moving Peaks Benchmark, we have confirmed the benefits of our strategies.

Models and Solutions for Truck and Trailer Routing Problems: An Overview

This paper is a survey about of the Truck and Trailer Routing Problem. The Truck and Trailer Routing Problem is an extension of the well-known Vehicle Routing Problem. Defined recently, this problem consists in designing the optimal set of routes for fleet of vehicles (trucks and trailers) in order to serve a given set of geographically dispersed customers. Since TTRP itself is a very difficult combinatorial optimization problem are usually tackled by metaheuristics. The interest in Truck and Trailer Routing Problem is motivated by its practical relevance as well as by its considerable difficulty. The goal of this paper is to show a study on the TTRP and the metaheuristics used for to solve it.

A software tool for assisting experimentation in dynamic environments

In real world, many optimization problems are dynamic, which means that their model elements vary with time. These problems have received increasing attention over time, especially from the viewpoint of metaheuristics methods. In this context, experimentation is a crucial task because of the stochastic nature of both algorithms and problems. Currently, there are several technologies whose methods, problems, and performance measures can be implemented. However, in most of them, certain features that make the experimentation process easy are not present. Examples of such features are the statistical analysis of the results and a graphical user interface (GUI) that allows an easy management of the experimentation process. Bearing in mind these limitations, in the present work, we present DynOptLab, a software tool for experimental analysis in dynamic environments. DynOptLab has two main components: (1) an object-oriented framework to facilitate the implementation of new proposals and (2) a graphical user interface for the experiment management and the statistical analysis of the results. With the aim of verifying the benefits of DynOptLabs main features, a typical case study on experimentation in dynamic environments was carried out.

Soft Computing Techniques and Sustainability Science, an Introduction

Sustainability Science is a research field that seeks to understand the fundamental character of interactions between Nature and Society. Because of the high degree of complexity of the problems and challenges it faces, this field requires new methodological approaches, tools and techniques that enable decision-makers and stakeholders can evaluate and make decisions based on a wide range of uncertainty and little information. It is here where Soft Computing methodologies can play an important role in addressing many of these challenges. The inherent tolerance of uncertainty and imprecision, and the robustness of the techniques that make up this paradigm can help solve or reduce the impact of these problems. This chapter introduces this book as a a catalogue of the most successful Soft Computing methodologies applied to Sustainability Science.

Modelling the interrelation among software quality criteria using Computational Intelligence techniques

Software products quality assessment is a highly complex process, given the variety of criteria to consider. For a better understanding, they are organized in so-called software quality models. An important aspect of these models is their structural complexity, forming a hierarchical structure. At present, they have evolved towards the overlap and interrelation between these criteria. Modeling a general structure to represent quality models, extending the hierarchical taxonomies to form more complicated structures like graphs and taking into account the interrelation between criteria is the objective of this work. The proposed solution incorporates elements of Computational Intelligence, such as fuzzy logic, fuzzy linguistic modeling and the use of fuzzy cognitive maps (FCM). The application of this proposal in a real-world case shows that it is an operative solution, reliable, precise and of easy interpretation for its application in the industry.

An Approach to Fault Diagnosis Using Fuzzy Clustering Techniques

In this paper a novel approach to design data driven based fault diagnosis systems using fuzzy clustering techniques is presented. In the proposal, the data was first pre-processed using the Noise Clustering algorithm. This permits to eliminate outliers and reduce the confusion as a first part of the classification process. Secondly, the Kernel Fuzzy C-means algorithm was used to achieve greater separability among the classes, and reduce the classification errors. Finally, it can be implemented a step for optimizing the parameters of the NC and KFCM algorithms. The proposed approach was validated using the iris benchmark data sets. The obtained results indicate the feasibility of the proposal.

Solving Regression Analysis by Fuzzy Quadratic Programming

Regression analysis, which includes any techniques for modeling and analyzing several variables, is a statistical tool that focuses in finding a relationship between a dependent variable and one or more independent variables. When this relationship is found, some values of parameters are determined which help a function to best fit in a set of data observations. In regression analysis, it is also interesting to characterize the variation of the depend variable around the independent ones. A regression problem can be formulated as a mathematical programming problem, where the objective is to minimize the difference between the estimated values and the observed values. This proposal provides a fuzzy solution to the problem that involves all particular -punctual- solutions provided by other methods. To clarify the above developments, a numerical example about the price mechanism of prefabricated houses is analyzed.