Algorithm Portfolios: Advances, Applications, and Challenges (SpringerBriefs in Optimization)

✍ Scribed by Dimitris Souravlias, Konstantinos E. Parsopoulos, Ilias S. Kotsireas, Panos M. Pardalos

• Publisher: Springer
• Year: 2021
• Tongue: English
• Leaves: 98
• Edition: 1st ed. 2021
• Category: Library

✦ Synopsis

This book covers algorithm portfolios, multi-method schemes that harness optimization algorithms into a joint framework to solve optimization problems. It is expected to be a primary reference point for researchers and doctoral students in relevant domains that seek a quick exposure to the field. The presentation focuses primarily on the applicability of the methods and the non-expert reader will find this book useful for starting designing and implementing algorithm portfolios. The book familiarizes the reader with algorithm portfolios through current advances, applications, and open problems. Fundamental issues in building effective and efficient algorithm portfolios such as selection of constituent algorithms, allocation of computational resources, interaction between algorithms and parallelism vs. sequential implementations are discussed. Several new applications are analyzed and insights on the underlying algorithmic designs are provided. Future directions, new challenges, and open problems in the design of algorithm portfolios and applications are explored to further motivate research in this field.

✦ Table of Contents

Preface
Suggested Audience and Outline of the Book
Acknowledgments
Contents
1 Metaheuristic Optimization Algorithms
1.1 Introduction
1.2 Trajectory-Based Metaheuristics
1.2.1 Tabu Search
1.2.2 Variable Neighborhood Search
1.2.3 Iterated Local Search
1.3 Population-Based Metaheuristics
1.3.1 Particle Swarm Optimization
1.3.2 Differential Evolution
1.3.3 Enhanced Differential Evolution
1.4 Synopsis
2 Algorithm Portfolios
2.1 Basics of Algorithm Portfolios
2.2 Design Challenges
2.3 Recent Developments
2.4 Synopsis
3 Selection of Constituent Algorithms
3.1 Algorithm Selection Problem
3.2 Algorithm Selection Methods
3.2.1 Feature-Based Selection
3.2.2 Statistical Selection
3.3 Synopsis
4 Allocation of Computation Resources
4.1 General Resource Allocation Problem
4.2 Online Resource Allocation
4.2.1 Algorithm Portfolio with Trading-Based ResourceAllocation
4.2.2 Algorithm Portfolio Based on Performance Forecasting
4.2.3 Adaptive Online Time Allocation Framework
4.3 Synopsis
5 Sequential and Parallel Models
5.1 Sequential Models
5.2 Parallel Models
5.3 Synopsis
6 Recent Applications
6.1 Application in Detection of Combinatorial Matrices
6.1.1 Problem Formulation
6.1.2 Known Existence Results
6.1.3 23 Years of Strassler's Table: 1997–2020
6.1.4 Solving the Problem Through Metaheuristics
6.1.5 First Algorithm Portfolio Approach
6.1.6 Second Algorithm Portfolio Approach
6.2 Application in Production Planning
6.2.1 Problem Formulation
6.2.2 Application of Algorithm Portfolios
6.3 Application in Humanitarian Logistics
6.3.1 Problem Formulation
6.3.2 Application of Algorithm Portfolio
6.4 Synopsis
7 Epilogue
References