Multi-Objective Particle Swarm Optimizers: A Survey of the State-of-the-Art

International Journal of Computational Intelligence Research, Journal Year: 2006, Volume and Issue: 2(3)

Published: Jan. 1, 2006

The success of the Particle Swarm Optimization (PSO) algorithm as a single-objective optimizer (mainly when dealing with continuous search spaces) has motivated researchers to extend use this bio-inspired technique other areas.One them is multi-objective optimization.Despite fact that first proposal Multi-Objective Optimizer (MOPSO) over six years old, considerable number algorithms have been proposed since then.This paper presents comprehensive review various MOPSOs reported in specialized literature.As part review, we include classification approaches, and identify main features each proposal.In last paper, list some topics within field consider promising areas future research.Baumgartner et al. [6] fully connected no Lexicographic ordering Hu Eberhart [24] ring yes rnd(0.5,1.0) [25] Sub-Population approaches Parsopoulos [49] Chow Tsui [8] Pareto-Based Moore Chapman [41] dominance Ray Liew [53] density estimator Fieldsend Singh [21] & closeness Coello [11, 12] solutions Toscano [66] randomly Srinivasan Hou [61] niche count

Language: Английский

---

Granular computing: an introduction

Published: Nov. 13, 2002

The study is concerned with the fundamentals of granular computing. Granular computing, as name itself stipulates, deals representing information in form some aggregates (that embrace a number individual entities) and their ensuing processing. We elaborate on rationale behind Next, formal frameworks granulation are discussed including several alternatives such fuzzy sets, interval analysis, rough probability. notion granularity defined quantified. A design agenda computing formulated key problems raised. architectures also an objective delineating fundamental algorithmic, conceptual challenges. It shown that use granules different size (granularity) lends to general pyramid role encoding decoding mechanisms visible this setting detail, along particular solutions. raise issue interoperability environments. intent paper put entire area certain perspective while not moving into specific algorithmic details.

Language: Английский

---

Self-adaptive Differential Evolution

Lecture notes in computer science, Journal Year: 2005, Volume and Issue: unknown, P. 192 - 199

Published: Jan. 1, 2005

Language: Английский

---

Efficient Population Utilization Strategy for Particle Swarm Optimizer

IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics), Journal Year: 2008, Volume and Issue: 39(2), P. 444 - 456

Published: Dec. 18, 2008

The particle swarm optimizer (PSO) is a population-based optimization technique that can be applied to wide range of problems. This paper presents variation on the traditional PSO algorithm, called efficient population utilization strategy for (EPUS-PSO), adopting manager significantly improve efficiency PSO. achieved by using variable particles in swarms enhance searching ability and drive more efficiently. Moreover, sharing principals are constructed stop from falling into local minimum make global optimal solution easier found particles. Experiments were conducted unimodal multimodal test functions such as Quadric, Griewanks, Rastrigin, Ackley, Weierstrass, with without coordinate rotation. results show good performance EPUS-PSO solving most benchmark problems compared other recent variants

Language: Английский

---

Tabu search for attribute reduction in rough set theory

Soft Computing, Journal Year: 2007, Volume and Issue: 12(9), P. 909 - 918

Published: Nov. 21, 2007

Language: Английский

---

Granular clustering: a granular signature of data

IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics), Journal Year: 2002, Volume and Issue: 32(2), P. 212 - 224

Published: April 1, 2002

The study is devoted to a granular analysis of data. We develop new clustering algorithm that organizes findings about data in the form collection information granules-hyperboxes. carried out here an example granulation mechanism. discuss compatibility measure guiding construction (growth) clusters and explain rationale behind their development. promotes mining way problem solving by emphasizing transparency results (hyperboxes). number indexes describing hyperboxes expressing relationships between such granules. It also shown how resulting family granules concise descriptor structure data-a signature examine properties features (variables) occurring as they manifest setting Numerical experiments are based on two-dimensional (2-D) synthetic well multivariable Boston available WWW.

Language: Английский

---

An alternative differential evolution algorithm for global optimization

Journal of Advanced Research, Journal Year: 2011, Volume and Issue: 3(2), P. 149 - 165

Published: July 26, 2011

The purpose of this paper is to present a new and an alternative differential evolution (ADE) algorithm for solving unconstrained global optimization problems. In the algorithm, directed mutation rule introduced based on weighted difference vector between best worst individuals particular generation. combined with basic strategy through linear decreasing probability rule. This modification shown enhance local search ability DE increase convergence rate. Two scaling factors are as uniform random variables improve diversity population bias direction. Additionally, dynamic non-linear increased crossover scheme utilized balance exploration exploitation. Furthermore, modified Breeder Genetic Algorithm (BGA) merged avoid stagnation and/or premature convergence. Numerical experiments comparisons set well-known high dimensional benchmark functions indicate that improved outperforms superior other existing algorithms in terms final solution quality, success rate, robustness.

Language: Английский

---

It really does matter how you normalize the branch distance in search‐based software testing

Software Testing Verification and Reliability, Journal Year: 2011, Volume and Issue: 23(2), P. 119 - 147

Published: March 23, 2011

SUMMARY The use of search algorithms for test data generation has seen many successful results. For structural criteria like branch coverage, heuristics have been designed to help the search. most common heuristic is approach level (usually represented with an integer) reward cases whose executions get close (in control flow graph) target branch. To solve constraints predicates in graph, distance commonly employed. These two measures are linearly combined. Since more important, normalized , often range [0, 1]. In this paper, different types normalizing functions analyzed. analyses show that one usually employed literature several flaws. paper presents a function very simple and does not suffer from these limitations. Empirical analytical carried out compare functions. particular, their effect studied on used algorithms, such as Hill Climbing, Simulated Annealing Genetic Algorithms. Copyright © 2011 John Wiley & Sons, Ltd.

Language: Английский

---

The comparison of multi-objective particle swarm optimization and NSGA II algorithm: applications in centrifugal pumps

Engineering Optimization, Journal Year: 2011, Volume and Issue: 43(10), P. 1095 - 1113

Published: April 6, 2011

In the present study, multi-objective optimization of centrifugal pumps is performed in three steps. first step, efficiency (η) and required net positive suction head (NPSHr) a set are numerically investigated using commercial software. Two meta-models based on evolved group method data handling (GMDH) type neural networks obtained second step for modeling η NPSHr with respect to geometrical design variables. Finally, polynomial networks, particle swarm (MOPSO) used Pareto-based considering two conflicting objectives, NPSHr. The Pareto results MOPSO also compared those genetic algorithm (NSGA II). It shown that some interesting important relationships as useful optimal principles involved performance can be discovered by metamodels representing characteristics.

Language: Английский

---

Linguistic models and linguistic modeling

IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics), Journal Year: 1999, Volume and Issue: 29(6), P. 745 - 757

Published: Jan. 1, 1999

The study is concerned with a linguistic approach to the design of new category fuzzy (granular) models. In contrast numerically driven identification techniques, we concentrate on budding meaningful labels (granules) in space experimental data and forming ensuing model as web associations between such granules. As models are designed at level information granules generate results same granular rather than pure numeric format, refer them Furthermore, there no detailed estimation procedures involved construction carried out this way, their mode can be viewed that rapid prototyping. underlying algorithm used development utilizes an augmented version clustering technique (context-based clustering) centered around notion contexts-a collection sets or relations defined (more precisely input variables). provided contrasted standard modeling approaches commonly encountered literature. usefulness system discussed illustrated aid studies including both synthetic well some time series dealing traffic intensity over broadband telecommunication network.

Language: Английский

---