Metaheuristics for optimization problems in computer communications

doi:10.1016/j.comcom.2006.08.027

Computer Communications

Volume 30, Issue 4, 26 February 2007, Pages 656-669

https://doi.org/10.1016/j.comcom.2006.08.027 Get rights and content

Abstract

Recent years have witnessed huge advances in computer technology and communication networks, entailing hard optimization problems in areas such as network design and routing. Metaheuristics are general high-level procedures that coordinate simple heuristics and rules to find good approximate solutions to computationally difficult combinatorial optimization problems. They are among the most effective solution strategies for solving optimization problems in practice and have been applied to a very large variety of problems in telecommunications, computer communications, and network design and routing. In this paper, we review the principles associated with some of the main metaheuristics and we give templates for basic implementations of them: simulated annealing, tabu search, GRASP, VNS, genetic algorithms, and path-relinking. The main strategies underlying the development of parallel implementations of metaheuristics are also reviewed. Finally, we present an account of some successful applications of metaheuristics to optimization problems in telecommunications, computer communications, and network design and routing.

Introduction

Combinatorial optimization problems [20], [62] involve finding optimal solutions from a discrete set of feasible solutions . However, even with the advent of new computer technologies and parallel processing, many of these problems cannot be solved to optimality in reasonable computation times, due to their inner nature or to their size. Moreover, reaching optimal solutions is meaningless in many practical situations, since we are often dealing with rough simplifications of reality and the available data is not precise. The goal of approximate algorithms (or heuristics) is to quickly produce good approximate solutions, without necessarily providing any guarantee of solution quality.

Metaheuristics are general high-level procedures that coordinate simple heuristics and rules to find good (often optimal) approximate solutions to computationally difficult combinatorial optimization problems. Among them, we find simulated annealing, tabu search, GRASP, genetic algorithms, scatter search, VNS, ant colonies, and others. They are based on distinct paradigms and offer different mechanisms to escape from locally optimal solutions, contrarily to greedy algorithms or local search methods. Metaheuristics are among the most effective solution strategies for solving combinatorial optimization problems in practice and they have been applied to a very large variety of areas and situations. The customization (or instantiation) of some metaheuristic to a given problem yields a heuristic to the latter.

The main principles underlying the most widely used metaheuristics are reviewed in the next section. Templates of these main metaheuristics are also given. Parallel implementations of metaheuristics entail more robust algorithms and appear quite naturally as an effective approach to speedup the search for good solutions. Strategies for the parallelization of metaheuristics are reviewed in Section 3. Applications of metaheuristics to several problems in telecommunications, computer communications, network design, and network routing are reviewed in Section 4.

Section snippets

Metaheuristics and templates

In this section, we review the main principles and ideas involved with greedy algorithms, local search procedures, and five widely and successfully used metaheuristics: simulated annealing, tabu search, GRASP, VNS, and genetic algorithms. We also discuss the main ideas underlining path-relinking, an intensification or post-optimization technique that can be used together with different metaheuristics.

In terms of notation, we consider a combinatorial optimization problem formulated as to

Parallelization of metaheuristics

The computation times associated with the exploration of the solution space may be very large. With the rapid increase in the number of parallel computers, powerful workstations, and fast communication networks, parallel implementations of metaheuristics appear quite naturally as an effective approach to speedup the search for approximate solutions. Besides the accelerations obtained, the parallelization also allows solving larger problems or finding better solutions.

Cung et al. [25] reviewed

Applications in computer communications

Communication networks consist of nodes that can be computers, database repositories, instruments like tomography equipments or radio transmitters, connected by data transmission links such as copper cables, optical fibers, satellite and radio links. Their design involves making decisions on many issues like the number of nodes and their locations, routing paths, capacity installation, wavelength allocation, and frequency assignment. The main objective is often to obtain a least cost network

References (83)

T.A. Feo et al.
A probabilistic heuristic for a computationally difficult set covering problem
Operations Research Letters
(1989)
F. Glover
Future paths for integer programming and links to artificial intelligence
Computers and Operations Research
(1986)
N. Mladenović et al.
Variable neighbourhood search
Computers and Operations Research
(1997)
T.F. Noronha et al.
Routing and wavelength assignment by partition coloring
European Journal of Operational Research
(2006)
A.Y. Zomaya et al.
Some observations on using meta-heuristics for efficient location management in mobile computing networks
Journal of Parallel and Distributed Computing
(2003)
E. Aarts et al.
Selected topics in simulated annealing
E.H.L. Aarts et al.
Simulated annealing and Boltzmann machines: a stochastic approach to combinatorial optimization and neural computing
(1989)
E.H.L. Aarts et al.
Local Search in Combinatorial Optimization
(1997)
J. Abello et al.
On maximum clique problems in very large graphs
J. Abello et al.
Massive quasi-clique detection
Lecture Notes in Computer Science
(2002)

R.M. Aiex, M.G.C. Resende, C.C.Ribeiro, TTTPLOTS: A perl program to create time-to-target plots, Optimization Letters,...

R.M. Aiex et al.

Probability distribution of solution time in GRASP: An experimental investigation

Journal of Heuristics

(2002)

E. Alba et al.

Solving the error correcting code problem with parallel hybrid heuristics

E. Alba et al.

On the behavior of parallel genetic algorithms for optimal placement of antennae in telecommunications

International Journal of Foundations of Computer Science

(2005)

E. Alba, C. Cotta, F. Chicano, A.J. Nebro, Parallel evolutionary algorithms in telecommunications: Two case studies,...

A.C. Alvim et al.

Load balancing for the parallelization of the GRASP metaheuristic

E. Amaldi et al.

Planning UMTS base station location: optimization models with power control and algorithms

IEEE Transactions on Wireless Communications

(2003)

M. Armony et al.

Design of stacked self-healing rings using a genetic algorithm

Journal of Heuristics

(2000)

G. Ausiello et al.

Complexity and Approximation: Combinatorial Optimization Problems and Their Approximability Properties

(1999)

M.P. Bastos et al.

Reactive tabu search with path relinking for the Steiner problem in graphs

L.S. Buriol et al.

A hybrid genetic algorithm for the weight setting problem in OSPF/IS-IS routing

Networks

(2005)

S. Cahon et al.

ParadisEO: a framework for the reusable design of parallel and distributed metaheuristics

Journal of Heuristics

(2004)

S.A. Canuto et al.

Local search with perturbations for the prize-collecting Steiner tree problem in graphs

Networks

(2001)

D. Castelino et al.

A surrogate constraint tabu thresholding implementation for the frequency assignment problem

Annals of Operations Research

(1999)

W.J. Cook et al.

Combinatorial Optimization

(1998)

T.H. Cormen et al.

Introduction to Algorithms

(2001)

L.A. Cox et al.

Designing least-cost survivable wireless backhaul networks

Journal of Heuristics

(2000)

T.G. Crainic et al.

Cooperative parallel tabu search for capacitated network design

Journal of Heuristics

(2002)

T.G. Crainic et al.

Parallel asynchronous tabu search in multicommodity location-allocation with balancing requirements

Annals of Operations Research

(1995)

V.-D. Cung et al.

Strategies for the parallel implementation of metaheuristics

G. Dahl et al.

The 2-path network problem

Networks

(2004)

T.A. Feo et al.

Greedy randomized adaptive search procedures

Journal of Global Optimization

(1995)

P. Festa et al.

Randomized heuristics for the max-cut problem

Optimization Methods and Software

(2002)

P. Festa et al.

GRASP: An annotated bibliography

A. Fink et al.

Solving general ring network design problems by meta-heuristics

V. Gabrel et al.

A comparison of heuristics for the discrete cost multicommodity network optimization problem

Journal of Heuristics

(2003)

B. Gendron et al.

Diversification strategies in local search for a nonbifurcated network loading problem

European Journal of Operational Research

(2000)

A. Girard et al.

A tabu search algorithm for access network design

Annals of Operations Research

(2001)

F. Glover

Tabu search – Part I

ORSA Journal on Computing

(1989)

F. Glover

Tabu search – Part I

ORSA Journal on Computing

(1990)

F. Glover

Tabu search and adaptive memory programming – Advances, applications and challenges

Cited by (36)

A bio-inspired approach: Firefly algorithm for Multi-Depot Vehicle Routing Problem with Time Windows
2022, Computer Communications
Citation Excerpt :
Once a vehicle reaches a customer point before ETi, later than LTi and after the maximum route time, waiting time occurs [5]. The firefly algorithm is a metaheuristic [6] technique that was used effectively in a variety of optimization disciplines. An improved bio-inspired metaheuristic technique, namely firefly, is used for tackling MDVRP-TW which is considered both the earliest arrival time and latest arrival time (double-sided time window) in this work.
In this research, an improved bio-inspired metaheuristic named firefly is utilized to solve the Multi-Depot Vehicle Routing Problem with Time Windows (MDVRP-TW) and it is a generalization of Vehicle Routing Problem with Time Windows. This paper discusses the use of biological intelligence in analysis and optimization on MDVRP-TW. The objective of the MDVRP-TW is to minimize the overall distance, means of transportation within a specified time gap, along with other constraints such as limitation of vehicle capacity, restrictions of route duration length to satisfy the customer’s requirements. The routes must be designed where each customer is visited only once in a particular time gap with one vehicle and all the routes must originate and end at the same depot. The overall demand should not go beyond the vehicle’s capacity on each route. Multi-objective problems have attracted the attentions of many researchers in recent years, as they are more closely related to real-world situations. However, transportation is a significant factor for the current society, exhibiting and planning a complicated transport system was a very thought-provoking task. Many research studies have shown several successful metaheuristics and heuristic methodologies to solve MDVRP-TW in which, instead of a single depot, numerous depots with varying locations. In this work considered both the earliest time and latest arrival time window (double-sided time window). It is assumed that if an arrival time is more significant than the latest time, it will lead to an infeasible solution whereas, if an arrival time is lesser than the earliest time, then it will lead to a waiting time. However, due to its challenging task and multi-objective and limitations, there is still a space to solve it. In this work, a novel technique that combines an improved firefly algorithm (IFA), which uses different mechanisms, namely, inter-depot approach in assignment phase, Clarke & wright algorithm, and local search method in routing and scheduling phase were applied. The suggested technique has been evaluated using Cordeau standard benchmark instances and proved the technique’s novelty and feasibility. In a few circumstances, the modified firefly algorithm outperforms other algorithms, and optimal benchmark solution in terms of results and competence.
Computational resource and cost prediction service for scientific workflows in federated clouds
2021, Future Generation Computer Systems
Citation Excerpt :
But in general, it deals with the possibility of estimating an unexpected conditional value, being a technique to model and investigate the relationship between variables. In addition to machine learning and the statistical techniques there is the metaheuristics that are procedures to combine different heuristics [29,30]. They differ basically by the mechanisms used to escape from local maximum.
The use of cloud platforms has been widely encouraged in applications that require a lot of processing power and storage such as scientific workflows. However, users who operate such workflows are faced with a very large variety and quantity of available resources, specially considering cloud federation platforms, making difficult to choose efficiently. Thus, the design issue to operate such workflows is far from trivial resulting in a problem to scientific workflows’ users. In order to address this problem, we propose a provisioning service called CRCPs - Computational Resource and Cost Prediction service that measure user resources and report the runtime financial cost before starting the workflow execution. In addition, CRCPs allows users to choose between high-performance, low-budget executions, or set how much to pay and how long to finish the workflow in an automatic and transparent way. The results show the CRCPs adequacy to estimate resources, time and execution cost of two different bioinformatic workflows executed using BioNimbuZ federated cloud platform.
Modern meta-heuristics based on nonlinear physics processes: A review of models and design procedures
2016, Physics Reports
Citation Excerpt :
Finally, hardware design problems have been also tackled with meta-heuristics approaches. In [287] the performance of several meta-heuristics (SA, TS, GRASP, VNS, EAs, and path-relinking) in telecommunications, computer communications, network design and routing problems are reviewed. Also related to hardware, in [288] a review of evolutionary algorithms in control of systems is carried out.
Meta-heuristic algorithms are problem-solving methods which try to find good-enough solutions to very hard optimization problems, at a reasonable computation time, where classical approaches fail, or cannot even been applied. Many existing meta-heuristics approaches are nature-inspired techniques, which work by simulating or modeling different natural processes in a computer. Historically, many of the most successful meta-heuristic approaches have had a biological inspiration, such as evolutionary computation or swarm intelligence paradigms, but in the last few years new approaches based on nonlinear physics processes modeling have been proposed and applied with success. Non-linear physics processes, modeled as optimization algorithms, are able to produce completely new search procedures, with extremely effective exploration capabilities in many cases, which are able to outperform existing optimization approaches. In this paper we review the most important optimization algorithms based on nonlinear physics, how they have been constructed from specific modeling of a real phenomena, and also their novelty in terms of comparison with alternative existing algorithms for optimization. We first review important concepts on optimization problems, search spaces and problems’ difficulty. Then, the usefulness of heuristics and meta-heuristics approaches to face hard optimization problems is introduced, and some of the main existing classical versions of these algorithms are reviewed. The mathematical framework of different nonlinear physics processes is then introduced as a preparatory step to review in detail the most important meta-heuristics based on them. A discussion on the novelty of these approaches, their main computational implementation and design issues, and the evaluation of a novel meta-heuristic based on Strange Attractors mutation will be carried out to complete the review of these techniques. We also describe some of the most important application areas, in broad sense, of meta-heuristics, and describe free-accessible software frameworks which can be used to make easier the implementation of these algorithms.
A hybrid data mining GRASP with path-relinking
2013, Computers and Operations Research
Citation Excerpt :
The second contribution is the application of the path-relinking hybrid proposals to solve the 2-path network design problem (2PNDP). This problem has shown to be NP-hard and many applications of this problem can be found in the design of communication networks, in which paths with few edges are sought to enforce high reliability and small delays [32]. The greedy algorithm presented in [7] and the GRASP procedures with path-relinking proposed in [33] are recent and important proposals developed to solve this problem.
The exploration of hybrid metaheuristics—combination of metaheuristics with concepts and processes from other research areas—has been an important trend in combinatorial optimization research. An instance of this study is the hybrid version of the GRASP metaheuristic that incorporates a data mining process. Traditional GRASP is an iterative metaheuristic which returns the best solution reached over all iterations. In the hybrid GRASP proposal, after executing a significant number of iterations, the data mining process extracts patterns from an elite set of sub-optimal solutions for the optimization problem. These patterns present characteristics of near optimal solutions and can be used to guide the following GRASP iterations in the search through the combinatorial solution space. The hybrid data mining GRASP has been successfully applied for different combinatorial problems: the set packing problem, the maximum diversity problem, the server replication for reliable multicast problem and the p-median problem. In this work, we show that, not only the traditional GRASP, but also GRASP improved with the path-relinking heuristic—a memory-based intensification strategy—could benefit from exploring a data mining procedure. Computational experiments, comparing traditional GRASP with path-relinking and different path-relinking hybrid proposals, showed that employing the combination of path-relinking and data mining made the GRASP find better results in less computational time. Another contribution of this work is the application of the path-relinking hybrid proposal for the 2-path network design problem, which improved the state-of-the-art solutions for this problem.
Soil-structure interaction: Parameters identification using particle swarm optimization
2011, Computers and Structures
This paper deals with a numerical NDT method to identify mechanical parameters by inverse analysis using the particle swarm optimization algorithm. This method is using a finite element model of the structure to study in order to minimize the error between field data and data predicted by the FE model. The method is applied in the case of a laboratory experiment modelling a soil–structure interaction problem. Results stress the impact of the sensitivity of the solution to the accuracy of field measurements by simulating both different levels of measurement noise on data and using several placements of sensors.
QoS-based cooperative algorithm for integral multi-commodity flow problem
2011, Computer Communications
Citation Excerpt :
Also implementing a decision support system including a variety of meta-heuristic algorithms [20] for the problem addressed by this paper will be pursued in next research.
This paper treats with integral multi-commodity flow through a network. To enhance the Quality of Service (QoS) for channels, it is necessary to minimize delay and congestion. Decreasing the end-to-end delay and consumption of bandwidth across channels are dependent and may be considered in very complex mathematical equations. To capture with this problem, a multi-commodity flow model is introduced whose targets are minimizing delay and congestion in one model. The flow through the network such as packets, also needs to get integral values. A model covering these concepts, is NP-hard while it is very important to find transmission strategies in real-time. For this aim, we extend a cooperative algorithm including traditional mathematical programming such as path enumeration and a meta-heuristic algorithm such as genetic algorithm. To find integral solution satisfying demands of nodes, we generalize a hybrid genetic algorithm to assign the integral commodities where they are needed. In this hybrid algorithm, we use feasible encoding and try to keep feasibility of chromosomes over iterations. By considering some random networks, we show that the proposed algorithm yields reasonable results in a few number of iterations. Also, because this algorithm can be applied in a wide range of objective functions in terms of delay and congestion, it is possible to find some routs for each commodity with high QoS. Due to these outcomes, the presented model and algorithm can be utilized in a variety of application in computer networks and transportation systems to decrease the congestion and increase the usage of channels.

View all citing articles on Scopus

Celso Ribeiro is full Professor at the Department of Computer Science of Universidade Federal Fluminense, Brazil. He chaired the Departments of Electrical Engineering (1983–1987) and Computer Science (1993–1995) of the Catholic University of Rio de Janeiro. He has a bachelor degree in Electrical Engineering (Catholic University of Rio de Janeiro, 1976) and an M.Sc. degree in Systems Engineering (Federal University of Rio de Janeiro, 1978). He obtained his doctorate in Computer Science at the École Nationale Supérieure des Télécommunications (Paris, France) in 1983. His research is supported by the Brazilian Council of Scientific and Technological Development (CNPq) and by the Rio de Janeiro State Foundation for Research Support (FAPERJ). Professor Ribeiro acted as President of the Brazilian Operations Research Society (SOBRAPO, 1989–1990) and of the Latin-American Association of Operations Research Societies (ALIO, 1992–1994), and as Vice-President of the International Federation of Operational Research Societies (IFORS, 1998–2000). He was a visiting researcher at AT&T Labs Research, International Computer Science Institute (ICSI, Berkeley), École Polytechnique de Montréal, and Université de Versailles (France). He is the editor of four books and the author of more than one hundred papers in international journals and book chapters. Professor Ribeiro supervised 14 doctorate dissertations and 30 master of science theses.

Simone de Lima Martins graduated with an Electrical Engineering degree from PUC-Rio in 1984, and obtained an M.Sc. degree in 1988 and a Ph.D. in 1999, both in Computer Science from PUC-Rio. She worked as a researcher at the Scientific Center of IBM-Brazil in 1988–1989 and from 1991–1993 as a computer systems analyst at IBM-Brazil, developing activities for voice and data integration in local area networks and multimedia systems. From 1999 to 2001, she worked as a visiting researcher in the Informatics Department of PUC-Rio and from 2001 to 2002 at LNCC (National Laboratory for Scientific Computation), both in Brazil. Since 2002, she is lecturer at the Universidade Federal Fluminense, where she develops activities in metaheuristics applications, bioinformatics, and parallel processing. She participates in several research projects financed by Brazilian government agencies in metaheuristics applications, parallel processing, and computational grids.

Isabel Rosseti graduated with a Computer Engineering degree from UFES in 1995, and obtained an M.Sc. degree in 1998 and a Ph.D. in 2003, both in Computer Science from PUC-Rio. From 2003 to 2005, she worked as a visiting researcher in the Computer Science Department of Universidade Federal Fluminense and from 2005 to 2006 she worked at ONS (Brazilian Independent System Operator) as Engineer in the areas of power system planning and operation. At present, she is lecturer at the Universidade Federal Fluminense, where she develops activities in metaheuristics applications, bioinformatics, and parallel processing.

View full text

Metaheuristics for optimization problems in computer communications

Abstract

Introduction

Section snippets

Metaheuristics and templates

Parallelization of metaheuristics

Applications in computer communications

Operations Research Letters

Computers and Operations Research

Computers and Operations Research

European Journal of Operational Research

Journal of Parallel and Distributed Computing

Selected topics in simulated annealing

Simulated annealing and Boltzmann machines: a stochastic approach to combinatorial optimization and neural computing

Local Search in Combinatorial Optimization

On maximum clique problems in very large graphs

Massive quasi-clique detection

Lecture Notes in Computer Science

Probability distribution of solution time in GRASP: An experimental investigation

Journal of Heuristics

Solving the error correcting code problem with parallel hybrid heuristics

On the behavior of parallel genetic algorithms for optimal placement of antennae in telecommunications

International Journal of Foundations of Computer Science

Load balancing for the parallelization of the GRASP metaheuristic

Planning UMTS base station location: optimization models with power control and algorithms

IEEE Transactions on Wireless Communications

Design of stacked self-healing rings using a genetic algorithm

Journal of Heuristics

Complexity and Approximation: Combinatorial Optimization Problems and Their Approximability Properties

Reactive tabu search with path relinking for the Steiner problem in graphs

A hybrid genetic algorithm for the weight setting problem in OSPF/IS-IS routing

Networks

ParadisEO: a framework for the reusable design of parallel and distributed metaheuristics

Journal of Heuristics

Local search with perturbations for the prize-collecting Steiner tree problem in graphs

Networks

A surrogate constraint tabu thresholding implementation for the frequency assignment problem

Annals of Operations Research

Combinatorial Optimization

Introduction to Algorithms

Designing least-cost survivable wireless backhaul networks

Journal of Heuristics

Cooperative parallel tabu search for capacitated network design

Journal of Heuristics

Parallel asynchronous tabu search in multicommodity location-allocation with balancing requirements

Annals of Operations Research

Strategies for the parallel implementation of metaheuristics

The 2-path network problem

Networks

Greedy randomized adaptive search procedures

Journal of Global Optimization

Randomized heuristics for the max-cut problem

Optimization Methods and Software

GRASP: An annotated bibliography

Solving general ring network design problems by meta-heuristics

A comparison of heuristics for the discrete cost multicommodity network optimization problem

Journal of Heuristics

Diversification strategies in local search for a nonbifurcated network loading problem

European Journal of Operational Research

A tabu search algorithm for access network design

Annals of Operations Research

Tabu search – Part I

ORSA Journal on Computing

Tabu search – Part I

ORSA Journal on Computing

Tabu search and adaptive memory programming – Advances, applications and challenges