Download Applications of Evolutionary Computation: EvoApplications 2010: EvoCOMNET, EvoENVIRONMENT, EvoFIN, EvoMUSART, and EvoTRANSLOG, Istanbul, Turkey, April 7-9, 2010, Proceedings, Part II PDF

TitleApplications of Evolutionary Computation: EvoApplications 2010: EvoCOMNET, EvoENVIRONMENT, EvoFIN, EvoMUSART, and EvoTRANSLOG, Istanbul, Turkey, April 7-9, 2010, Proceedings, Part II
Author
LanguageEnglish
File Size9.5 MB
Total Pages504
Table of Contents
                            Cover
Front matter
1. Detection of DDoS Attacks via an Artificial Immune System-Inspired Multiobjective Evolutionary Algorithm
	Detection of DDoS Attacks via an Artificial Immune System-Inspired Multiobjective Evolutionary Algorithm
		Introduction
		Background
			Intrusion Detection Systems (IDS)
			Distributed Denial of Service Attack (DDoS)
			Datasets
		JREMISA (Java REtrovirus-inspired Multiobjective Immune System Algorithm)
			Representation of Antigens and Antibodies
			Immune Algorithm
		Proposed Improvements on jREMISA
		Experiments
			Test Designs
			Results
		Conclusion
		References
2. Performance Evaluation of an Artificial Neural Network-Based Adaptive Antenna Array System
	Performance Evaluation of an Artificial Neural Network-Based Adaptive Antenna Array System
		Introduction
		Principle AAA System Configuration
		The Smart Neural AAA System Model
			Data Processing Unit
			The ANNP Unit
		Discussion of Simulation Results
		Conclusions
		References
3. Automatic Parameter Tuning with Metaheuristics of the AODV Routing Protocol for Vehicular Ad-Hoc Networks
	Automatic Parameter Tuning with Metaheuristics of the AODV Routing Protocol for Vehicular Ad-Hoc Networks
		Introduction
		AODV Parameter Tuning
		Optimization Strategy
		VANET Scenario and Mobility Models
		Experiments
			Parameter Settings of the Optimization Algorithms
			Simulation Results and Comparisons
			QoS Analysis
		Conclusions
4. WiMAX Network Planning Using Adaptive-Population-Size Genetic Algorithm
	WiMAX Network Planning Using Adaptive-Population-Size Genetic Algorithm
		Introduction
		Background
			Problem Formulation
			Related Works
		Adaptive-Population-Size Genetic Algorithm
			Individual Representation
			Evolution Framework
			Adaptive Population Size Adjustment
			Evolutionary Operations
		Simulation
			Network Layouts and Algorithm Configuration
			Results
		Concluding Remarks and Future Research
5. Markov Chain Models for Genetic Algorithm Based Topology Control in MANETs
	Markov Chain Models for Genetic Algorithm Based Topology Control in MANETs
		Introduction
		Our Forced-Based GA
		Ergodic Homogeneous Finite Markov Chains
			Convergence of Ergodic Homogeneous Finite Markov Chain
		Simulation Experiments of Convergence for Our Forced-Based GA
		Conclusion and Future Work
6. Particle Swarm Optimization for Coverage Maximization and Energy Conservation in Wireless Sensor Networks
	Particle Swarm Optimization for Coverage Maximization and Energy Conservation in Wireless Sensor Networks
		Introduction
		Particle Swarm Optimization
		PSO for WSN Coverage Optimization
			Particle Encoding
			Fitness Function
		Results and Discussion
		Conclusion
		References
7. Efficient Load Balancing for a Resilient Packet Ring Using Artificial Bee Colony
	Efficient Load Balancing for a Resilient Packet Ring Using Artificial Bee Colony
		Introduction
		Problem Definition
		Proposed ABC
		Results
		Conclusions
		References
8. TCP Modification Robust to Packet Reordering in Ant Routing Networks
	TCP Modification Robust to Packet Reordering in Ant Routing Networks
		Introduction
			Swarm Intelligence and Ant Routing
			The TCP Protocol
		The TCP Modification
			Packet Delay Modeling
			Delay Model Based TCP
		Experimental Results
			Network without Packet Loss
			Network with Packet Loss
		Conclusions
9. Solving the Physical Impairment Aware Routing and Wavelength Assignment Problem in Optical WDM Networks Using a Tabu Search Based Hyper-Heuristic Approach
	Solving the Physical Impairment Aware Routing and Wavelength Assignment Problem in Optical WDM Networks Using a Tabu Search Based Hyper-Heuristic Approach
		Introduction
		Literature Survey and Problem Definition
			Literature Survey
			Routing and Wavelength Assignment Problem
		Solution Approaches
			Hyper-Heuristics
			Proposed Method
		Experimental Design
		Results
		Conclusion and Future Work
10. A Generalized, Location-Based Model of Connections in Ad-Hoc Networks Improving the Performance of Ant Routing
	A Generalized, Location-Based Model of Connections in Ad-Hoc Networks Improving the Performance of Ant Routing
		Introduction
		Related Work
		Generalized Model of Connections
			Ad-hoc network graph - nodes level.
				Ad-hoc network graph - locations level.
		AntHocGeo Algorithm
		Experimental Results
			Connections Stability
			The Overhead
			Overall Performance
		Conclusions
11. Using Code Bloat to Obfuscate Evolved Network Traffic
	Using Code Bloat to Obfuscate Evolved Network Traffic
		Introduction
		Background
			Port Scans
			TCP/IP Packets
			Code Bloat
		Evolutionary Model
			Instruction Set
			Fitness Evaluation
			Evolutionary Model
		Experiments and Results
			Analysis
		Conclusion and Future Work
12. ABC Supported Handoff Decision Scheme Based on Population Migration
	ABC Supported Handoff Decision Scheme Based on Population Migration
		Introduction
		Model Description
			Application Type, QoS Requirement and Its Fuzzy Degree
			Access Network Model and Terminal Model
			Satisfaction Degree and Suitability Degree
			Gaming Analysis and Utility Calculation
			Mathematical Model
		Algorithm Description
			Solution Encoding and Its Attraction Force Function
			Algorithm Procedure
		Simulation Implementation and Performance Evaluation
		Conclusion
		References
13. A Hyper-Heuristic Approach for the Unit Commitment Problem
	A Hyper-Heuristic Approach for the Unit Commitment Problem
		Introduction
		The Unit Commitment Problem
		Hyper-Heuristics
		Related Work on the UCP
		Proposed Approach
		Experiments
			Parameter Settings
			Experimental Results
		Conclusion
		References
14. Application of Genetic Programming Classification in an Industrial Process Resulting in Greenhouse Gas Emission Reductions
	Application of Genetic Programming Classification in an Industrial Process Resulting in Greenhouse Gas Emission Reductions
		Introduction
		Data and Methods
		Experiments and Results
			The CART Model
			GP Evolved Pure Classification Rules (Class-GP)
			GP Evolved Regression Function (Reg-GP)
		Discussion
		Conclusions
15. Influence of Topology and Payload on CO 2  Optimised Vehicle Routing
	Influence of Topology and Payload on CO2 Optimised Vehicle Routing
		Introduction
		Related Work
		CO2 Emission Modelling
		Experimental Approach
		Evidence
		Conclusions
16. Start-Up Optimisation of a Combined Cycle Power Plant with Multiobjective Evolutionary Algorithms
	Start-Up Optimisation of a Combined Cycle Power Plant with Multiobjective Evolutionary Algorithms
		Introduction
		Multiobjective Optimisation
			Multiobjective Evolutionary Algorithms
		Start-Up Optimisation of a Combined Cycle Power Plant
		Experimentations
		Results
			Discussion
		Conclusion and Future Work
17. A Study of Nature-Inspired Methods for Financial Trend Reversal Detection
	A Study of Nature-Inspired Methods for Financial Trend Reversal Detection
		Introduction
		Problem Description
		Detector Set Definition and Nature-Inspired Methodologies
			Particle Swarm Optimization
			Negative Selection
		Experiments and Results
			Discussion
		Conclusion and Future Work
18. Outperforming Buy-and-Hold with Evolved Technical Trading Rules: Daily, Weekly and Monthly Trading
	Outperforming Buy-and-Hold with Evolved Technical Trading Rules: Daily, Weekly and Monthly Trading
		Introduction
		Evolving Robust Trading Rules: The Modified AK/BS Approach
			Overview
			Function and Terminal Sets
			The Fitness Function
			Operators and Initialization
		Experiments
			GP Parameters, Data Periods and Consistency-of Performance Periods
			Results
		Concluding Summary and Discussion
		References
19. Evolutionary Multi-stage Financial Scenario Tree Generation
	Evolutionary Multi-stage Financial Scenario Tree Generation
		Introduction
		Multi-stage Scenario Tree Generation
		Evolutionary Multi-stage Scenario Tree Generation
		Numerical Results
		Conclusion
20. Evolving Dynamic Trade Execution Strategies Using Grammatical Evolution
	Evolving Dynamic Trade Execution Strategies Using Grammatical Evolution
		Introduction
		Background
		Evolving Dynamic Trade Execution Strategies
			Information Indicators
			Grammar of Grammatical Evolution Algorithm
			Performance Evaluation
		Simulating an Artificial Market
		Results
		Conclusions and Future Work
21. Modesty Is the Best Policy: Automatic Discovery of Viable Forecasting Goals in Financial Data
	Modesty Is the Best Policy: Automatic Discovery of Viable Forecasting Goals in Financial Data
		Introduction
		Background
		The Hybrid Forecasting System
			Representation.
			Fitness Evaluation
				Calculating the Predictability for each Genome.
				Calculating Profitability for each Genome.
				Calculating the Dominance Count for each Genome.
			Sexual Operations
			Additional Considerations
		Experimental Setup
			Data preparation
			Comparison Approaches
		Results
		Conclusions
22. Threshold Recurrent Reinforcement Learning Model for Automated Trading
	Threshold Recurrent Reinforcement Learning Model for Automated Trading
		Introduction
		Model Description
			Threshold Recurrent Reinforcement Learning
			Differential Sharpe Ratio for Online Learning
		Experiments
			Setup
			Data
			Results and Discussion
		Conclusion
23. Active Portfolio Management from a Fuzzy Multi-objective Programming Perspective
	Active Portfolio Management from a Fuzzy Multi-objective Programming Perspective
		Introduction
		Index Tracking with Cardinality Constraints
		Nature-Inspired Optimisation Algorithms
		Traditional vs. Fuzzy Enhanced Indexation
		Empirical Study: Actively Reproducing the Dow Jones Industrial Average Index
			Sample Data and Experimental Design
		Discussion - Further Research
24. Evolutionary Monte Carlo Based Techniques for First Passage Time Problems in Credit Risk and Other Applications in Finance
	Evolutionary Monte Carlo Based Techniques for First Passage Time Problems in Credit Risk and Other Applications in Finance
		Introduction
		First Passage Time in Credit Risk Models
		Multivariate Jump-Diffusion Processes and Monte Carlo Simulations
		Density Functions, Default Rates, and Correlated Default
		Conclusion
25. Calibrating the Heston Model with Differential Evolution
	Calibrating the Heston Model with Differential Evolution
		Introduction
		Pricing with the Characteristic Function
			Black–Scholes–Merton
			The Heston Model
			Integration Schemes
		Calibrating the Model Parameters
			Differential Evolution
			Calibrating the Heston Model
		Conclusion
26. Evolving Trading Rule-Based Policies
	Evolving Trading Rule-Based Policies
		Introduction
			Structure of Paper
		Rule-Based Policies
		Grammatical Representation
		Evolution of Trading Policies
			Data Review
			Methodology
		Results
			Example Evolved Policy
		Conclusion and Future Work
27. Evolving Artistic Styles through Visual Dialogues
	Evolving Artistic Styles through Visual Dialogues
敳敲癥搠䁤 㴀 ⨀䁬整䁴潫敮 ⴀ㈀洀
		Introduction
		Background and Motivation
		A Description of I3
			The Top-Level Behavior
			The Generator Component
			The Analyzer Component
		I3 Experience
		Summary and Future Work
28. Graph-Based Evolution of Visual Languages
	Graph-Based Evolution of Visual Languages
		Introduction
		Context Free
		Evolutionary Context Free Art
			Crossover Operator
			Mutation Operators
		Experimentation
			Fitness Functions
			Experimental Results
		Conclusions and Future Work
29. Refinement Techniques for Animated Evolutionary Photomosaics Using Limited Tile Collections
	Refinement Techniques for Animated Evolutionary Photomosaics Using Limited Tile Collections
		Introduction
		Related Work
		Refinement Strategies in Photomosaic Generation
			Colour Adjustment
			Tile Size Variation
			Fitness Evaluation
		Results and Discussion
			Colour Adjustment
			Tile Size Variation
		Conclusions
30. Generative Art and Evolutionary Refinement
	Generative Art and Evolutionary Refinement
		Introduction
		Our Cellular Morphogenesis Evolutionary Art System
		Evolutionary Exploration Using Fitness Functions
		Analysis of the Source Material
		Evolutionary Refinement of the Source Material
		Conclusion
31. Aesthetic Learning in an Interactive Evolutionary Art System
	Aesthetic Learning in an Interactive Evolutionary Art System
		Introduction
		Features for Measuring Aesthetics
			Image Complexity Estimation
			Image Order Estimation
		Experimental Results
			Experimental Setup
				Genetic Programming.
				Mutation Operators.
				Evolutionary and Learning Process.
			Aesthetic Learning
				Accuracy of Prediction.
				Decision Tree.
				Generation Results.
			Validation Study
		Conclusions and Future Work
32. Comparing Aesthetic Measures for Evolutionary Art
	Comparing Aesthetic Measuresfor Evolutionary Art
		Introduction
			Research Question
		Evolutionary Art
			Four Aesthetic Measures
		Arabitat: The Art Habitat
		Experiments
			Results
			Cross Evaluation
		Conclusions
		Future Work
33. The Problem with Evolutionary Art Is ...
	The Problem with Evolutionary Art Is …
		Introduction
		The Problem of Fitness Functions for Evolutionary Art
			Interactive Evolutionary Computing
			Computational Aesthetic Evaluation
			Hybrid Aesthetic Evaluation
			The Future of Aesthetic Evaluation for Evolutionary Art
		The Problem of Genetic Representation and Innovation
			Complexification in Nature and Genetic Representation
		The Problem of Art Theory for Evolutionary Art
			Evolutionary Art Theory and Truth to Process
		References
34. Learning to Dance through Interactive Evolution
	Learning to Dance through Interactive Evolution
		Introduction
		Background
		Approach
			ANN Inputs
			Audio Processing
			ANN Outputs
			ANN Training
		Experiments and Results
			Dancing to MIDI
			Dancing to Raw Audio
		Discussion
		Conclusion
35. Jive: A Generative, Interactive, Virtual, Evolutionary Music System
	Jive: A Generative, Interactive, Virtual, Evolutionary Music System
		Introduction
		Previous Work
		The Basic Jive System
			Generative
			Interactive
			Virtual
			Evolutionary
		Results and Refinements
		Discussion
		Conclusions and Future Work
36. A Neural Network for Bass Functional Harmonization
	A Neural Network for Bass Functional Harmonization
		Introduction
		The Neural Networks
		Training, Validation and Test Results
		Conclusions
37. Combining Musical Constraints with Markov Transition Probabilities to Improve the Generation of Creative Musical Structures
	Combining Musical Constraints with Markov Transition Probabilities to Improve the Generation of Creative Musical Structures
		Markov Chains and Music
		Musical Constraints
		Combining Musical Constraints and Transition Probabilities
			Tackling Drift and the End Point Problems with Stochastic Optimization
			Simulated Thermal Annealing
			Annealing Results
		Conclusions
		References
38. Dynamic Musical Orchestration Using Genetic Algorithms and a Spectro-Temporal Description of Musical Instruments
	Dynamic Musical Orchestration Using Genetic Algorithms and a Spectro-Temporal Description of Musical Instruments
		Introduction
		System Architecture
		Temporal Descriptor
			Existing Models
			Our Model
			Length Modification
				Modification by dilation.
				Modification by repetition.
			Results
		Experiments
		Conclusion and Future Work
39. Evolutionary Sound Synthesis: Rendering Spectrograms from Cellular Automata Histograms
	Evolutionary Sound Synthesis: Rendering Spectrograms from Cellular Automata Histograms
		Introduction
		The Multitype Voter Model
		Mapping Process: From Histograms to Spectrograms
		Features of the Multitype Voter Model Histogram Sequences
		Attacks and Releases
		Control
		Conclusion and Further Work
		References
40. Sound Agents
	Sound Agents
		Introduction
		Hardware Implementation of the Sound Agents System
		Swarm Intelligence
		A Declarative Language for Describing Agent Behaviors
			Goal Constraints and Local Search Constraint Solving
			Goal Constraints for Navigation
		Conclusion
		References
41. From Evolutionary Composition to Robotic Sonification
	From Evolutionary Composition to Robotic Sonification
		Introduction
		AURAL Architecture
			The OmniEye
			The Evolutionary Sound Interface
			Robotic Control
		Experiments
			Collective Behavior Affects Performance Control
		Conclusion
		References
42. Musical Composer Identification through Probabilistic and Feedforward Neural Networks
	Musical Composer Identification through Probabilistic and Feedforward Neural Networks
		Introduction
		Data Set and Data Extraction
		Classification Methods Tested
		Methodology and Experimental Results
		Discussion and Concluding Remarks
43. Using an Evolutionary Algorithm to Discover Low CO 2  Tours within a Travelling Salesman Problem
	Using an Evolutionary Algorithm to Discover Low CO2 Tours within a Travelling Salesman Problem
		Introduction and Motivation
		Previous Work
		Problem Description
			The Geographical Data Source
			Estimating Vehicle Emissions
			Emissions Calculations Using a Fuel Consumption Model
			Emissions Calculations Using a Simpler Model
		Experimental Method and Results
			Problem Instances
			The Evolutionary Algorithm Employed
			Experimental Method
			Results
		Conclusions and Future Work
44. A Genetic Algorithm for the Traveling Salesman Problem with Pickup and Delivery Using Depot Removal and Insertion Moves
	A Genetic Algorithm for the Traveling Salesman Problem
		Introduction
		An Evolutionary Approach for the TSPPD
			A Genetic Algorithm
			Tour Improvement Procedure
		Computational Experiments
		Conclusion
45. Fast Approximation Heuristics for Multi-Objective Vehicle Routing Problems
	Fast Approximation Heuristics for Multi-Objective Vehicle Routing Problems
		Introduction
		Problem Statement
		Solution Approach
			Encoding of Alternatives
			Constructive Phase: A Multi-objective Savings Heuristic
			Iterative Phase: Population-Based Multi-operator Search
		Experimental Investigation
			Benchmark Data
			Experiments and Results
		Conclusions
46. Particle Swarm Optimization and an Agent-Based Algorithm for a Problem of Staff Scheduling
	Particle Swarm Optimization and an Agent-Based Algorithm for a Problem of Staff Scheduling
		Introduction
		A Real-World Problem from Logistics
		Related Work
		PSO and Artificial Agent Approach
			PSO for This Application
			Artificial Agents for This Application
		Results and Discussion
		Conclusion and Future Work
47. A Math-Heuristic for the Multi-Level Capacitated Lot Sizing Problem with Carryover
	A Math-Heuristic for the Multi-Level Capacitated Lot Sizing Problem with Carryover
		Introduction
		A Formal Model for the MLCLSP-CO
		General Idea and Algorithm
		Incumbent Solution Generation: A Metaheuristic Scheme
		Computational Results
		Conclusions
Back matter
                        
Document Text Contents
Page 2

Lecture Notes in Computer Science 6025
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen

Editorial Board

David Hutchison
Lancaster University, UK

Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA

Josef Kittler
University of Surrey, Guildford, UK

Jon M. Kleinberg
Cornell University, Ithaca, NY, USA

Alfred Kobsa
University of California, Irvine, CA, USA

Friedemann Mattern
ETH Zurich, Switzerland

John C. Mitchell
Stanford University, CA, USA

Moni Naor
Weizmann Institute of Science, Rehovot, Israel

Oscar Nierstrasz
University of Bern, Switzerland

C. Pandu Rangan
Indian Institute of Technology, Madras, India

Bernhard Steffen
TU Dortmund University, Germany

Madhu Sudan
Microsoft Research, Cambridge, MA, USA

Demetri Terzopoulos
University of California, Los Angeles, CA, USA

Doug Tygar
University of California, Berkeley, CA, USA

Gerhard Weikum
Max-Planck Institute of Computer Science, Saarbruecken, Germany

Page 252

Active Portfolio Management 223

Nowadays, there exist a plethora of methodologies for structuring index track-
ing portfolios with cardinality constraints (some references are given in following
sections; see also [1] for a comprehensive survey). Despite the numerous applica-
tions of passive portfolio selection, active tracking formulations - especially those
incorporating cardinality constraints - have not yet attracted full attention. In
[8] we consider three alternative formulations of active portfolio management,
based on the traditional mean-variance view of portfolio selection. In this paper,
we expand upon this issue by making an attempt to incorporate non-standard
objectives related to portfolio performance. These focus, for example, on the
probability that the index portfolio delivers positive return relatively to the
benchmark or on restricting the total risk of the investment strategy. Such tar-
gets/constraints on portfolio performance can be effectively formulated within
the framework of fuzzy mathematical multi-objective programming.

The rest of the article is structured as follows: enhanced index tracking with
cardinality constraints is studied in section 2, while section 3 discusses the
nature-inspired optimisation heuristics employed in our study to solve active
portfolio formulations. Section 4 details an alternative, “fuzzy”, conceptualisa-
tion of active portfolio management structured around approximate investment
targets and portfolio constraints. In section 5 we evaluate the performance of
fuzzy portfolios in terms of actively reproducing the American DJIA index. Sec-
tion 6 summarises the main findings and proposes future research directions.

2 Index Tracking with Cardinality Constraints

Consider the following investment problem whereby a fund manager has to de-
cide the best subset of K stocks (K < N) that can actively reproduce returns
on a benchmark index I as well as the appropriate percentage of capital wi that
should be invested in each stock i. The active portfolio optimisation problem can
be stated in its general form as maximise

w∈RN , s∈{0,1}N
f(w, s) subject to

∑N
i=1 wi = 1,

wl ≤ wi ≤ wu, i = 1, ..., N and
∑N

i=1 si ≤ K ≤ N , where f(w, s) is the port-
folio’s active objective, which is a function of weights w = (w1, w2, ..., wN )′, a
vector of binary variables s = (s1, s2, ..., sN )′ and sample data. Several choices
for f(., .) are later discussed in section 4. Each si is an indicator function that
takes the value 1 if asset i is included in the portfolio and 0 otherwise. All asset
weights sum to one, implying that the initial capital is fully invested, and the
maximum number of assets allowed in the portfolio should not exceed K ≤ N ,
which is the cardinality constraint. Depending on the cardinality of the portfolio
P , some of the wi’s may be zero and if an asset i is included then the fund man-
ager may impose a lower or upper limit on its weight, wl and wu, respectively,
the so-called floor and ceiling constraint.

The introduction of cardinality constraints significantly increases the com-
putational effort associated with deriving optimal portfolio allocations. In fact,
one ends up with a mixed nonlinear-integer programming problem, which even

Page 253

224 N.S. Thomaidis

for small values of N becomes a challenge for gradient-based optimisation tech-
niques. In this paper, we adopt a solution strategy that overcomes the difficulties
of handling integer constraints by introducing a transformation of the decision
variables (see [8,6] for details). The underlying idea is to solve an unconstrained
optimisation programme with decision variables (x1, ...., xN ) ∈ RN and use a
deterministic function to map the values of (x1, ...., xN ) onto a feasible portfolio
allocation (w1, w2, ..., wN ), where K out of N weights are zero.

3 Nature-Inspired Optimisation Algorithms

Many real-life optimisation problems in finance are considered “hard” due to
combinatorial explosion or the ruggedness of the optimisation landscape. Tradi-
tionally, practitioners approach these problems adopting techniques that make
use of relaxation and decomposition principles, such as branch & bound, dy-
namic programming and quadratic line-search. In the last decades, a number of
intelligent computational algorithms, such as simulated annealing, tabu search,
genetic algorithms, ant colonies and particle swarms, have been developed to
solve a wide range of practical optimisation problems.

In this study, we experiment with three popular nature-inspired computa-
tional schemes in the task of solving active portfolio optimisation problems. We
first present a trajectory-based strategy, namely simulated annealing, and then
introduce two evolutionary computational heuristics, genetic algorithms and par-
ticle swarm optimisation.

Simulated annealing (SA) took its name and inspiration from the annealing,
a technique used in metallurgy involving heating and controlled cooling of a
solid ([5]). The central idea is to start with some arbitrary solution and have
it modified by randomly generating a number (or else population) of new so-
lutions in its vicinity. To overcome the possibility of premature convergence to
local optima, the algorithm also accepts solutions that come with impairment,
yet with decreasing probability. Genetic algorithms (GAs) employ an alterna-
tive solution-search strategy by forcing a parallel exploration of the solution
space by means of several agents that interact with each other. Inspired by the
process of natural selection that drives biological evolution, a genetic algorithm
repeatedly modifies a population of solutions until it“evolves” towards a “fit”
generation ([7]). Among the numerous forms in which genetic algorithms ap-
pear in the literature, in this paper we adopt a simple version of the algorithm
that encodes solutions into vectors of real numbers and uses three genetic opera-
tors to move towards fitter populations: elitist selection, crossover and mutation.
Particle swarm optimisation (PSO) is another computational heuristic inspired
by the behaviour of biological flocks or swarms [3,4]. Instead of applying ge-
netic operators, PSO flows “particles” in the search space with a velocity and
direction that are dynamically adjusted according to each particle’s learning ex-
perience (the best solution detected by the particle in its own exploration) and
the swarm’s collective memory (the global best solution found by any particle
in the swarm).

Page 503

Author Index 475

Oh, Jae C. II-261
Olague, Gustavo I-344
Oliveto, Pietro Simone I-61
Öncan, Temel II-431
O’Neill, Michael I-161, II-192,

II-251, II-341
Oranchak, David I-181

Pacut, Andrzej II-71, II-91
Pantano, Pietro I-211
Pasquet, Olivier II-391
Pasquier, Philippe I-131
Payne, Joshua L. I-41
Peña, José-Maŕıa I-422
Pizzo, Christian II-41
Piazzolla, Alessandro I-320
Pizzuti, Stefano II-151
Pospichal, Petr I-442
Protopapas, Mattheos K. I-191

Qin, Peiyu II-111
Quattrone, Aldo I-211

Ramirez, Adriana II-462
Ramtohul, Tikesh II-212
Rees, Jonathan I-312
Reynoso-Meza, Gilberto I-532
Richter, Hendrik I-552
Rimmel, Arpad I-201
Rocchisani, Jean-Marie I-292
Rolet, Philippe I-592
Romero, Juan II-271
Runarsson, Thomas Philip I-121
Ryan, Conor II-202

Şahin, Cem Şafak II-41
Sánchez, Ernesto I-11, I-412
Sánchez, Pablo Garćıa I-171
Sánchez-Pérez, Juan Manuel II-61
Sanchis, Javier I-532
Sarasola, Briseida I-572
Sbalzarini, Ivo F. I-432
Schettini, Raimondo I-282
Schumann, Enrico II-242
Schwarz, Josef I-442
Scionti, Alberto I-412
Scott, Cathy II-141, II-421
Seo, Kisung I-352, I-381
Serquera, Jaime II-381

Shao, Jianhua II-341
Sharabati, Anas I-361
Shukla, Pradyumn Kumar I-21
Silva, Sara I-272, II-131
Şima Uyar, A. II-1, II-81, II-121
Smit, S.K. I-542
Sorenson, Nathan I-131
Sossa, Humberto I-344
Squillero, Giovanni I-11, I-412
Staino, Andrea I-211
Stanley, Kenneth O. I-141, II-331
Stramandinoli, Francesca I-211
Süral, Haldun II-431
Swafford, John Mark I-161
Szeto, K.Y. I-151

Tamimi, Hashem I-361
Tenne, Yoel I-582
Tettamanzi, Andrea G.B. II-161
Teytaud, Fabien I-201, I-452
Teytaud, Olivier I-592
Thomaidis, Nikos S. II-222
Tirronen, Ville I-471
Togelius, Julian I-141
Tonda, Alberto I-11, I-412
Torre, Antonio II-351
Trucco, Emanuele I-241
Tsviliuk, Olena II-232

Urquhart, Neil II-141, II-421
Urrea, Elkin II-41
Uyar, M. Ümit II-41

Vanneschi, Leonardo I-282
Vasconcelos, Maria J. I-272
Vega-Rodŕıguez, Miguel Angel II-61
Vidal, Franck Patrick I-292
Villegas-Cortez, Juan I-344
Voß, Stefan II-462
Vrahatis, Michael N. II-411

Waldock, Antony I-461
Wang, Xingwei II-111
Watson, Richard A. I-1
Weber, Matthieu I-471
Wong, Ka-Chun I-481
Wong, Man-Hon I-481
Wu, Chun-Ho I-302
Wu, Degang I-151

Page 504

476 Author Index

Yang, Shengxiang I-491, I-562
Yannakakis, Georgios N. I-141
Yao, Xin I-61
Yayımlı, Ayşegül II-81
Yung, Kei-Leung I-302

Zaccagnino, Rocco II-351
Zajec, Edward II-261
Zhang, Di II-232
Zhang, Mengjie II-51
Zincir-Heywood, Nur II-101

Similer Documents