The aim of this book is to introduce Harmony Search algorithm in the context of solving real life problems. Harmony Search (HS) is a musician’s behavior inspired metaheuristic algorithm developed in 2001, though it is a relatively new meta heuristic algorithm, its effectiveness and advantages have been demonstrated in various applications like traffic routing, multi objective optimization, design of municipal water distribution networks, load dispatch problem in electrical engineering, rostering problems, clustering, structural design, classification and feature selection to name a few.
Optimization is the process of finding the best alternate solution among a given set of solutions under some given constraints. The process of finding the maximum or minimum possible value, which a function can attain in its domain, is known as optimization. One of the most striking trends that emerged in the optimization field is the simulation of natural processes as efficient global search methods.
The natural processes or phenomena are firstly analyzed mathematically and then coded as computer programs for solving complex nonlinear real-world problems. The resulting methods are called Nature Inspired Algorithms that can often outperform classic methods. The advantages of these methods are their ability to solve various standard or application-based problems successfully without any prior knowledge of the problem space. Moreover, these algorithms are more likely to obtain the global optima of a given problem. They do not require any continuity and differentiability of the objective functions.
Also, they work on a randomly generated population of solutions instead of one solution. They are easy to program and can be easily implemented on a computer. Some of the examples of Nature Inspired Optimization Techniques are Genetic Algorithm, Particle Swarm Optimization, Artificial Bee Colony Optimization and Ant Colony Optimization.
Table of Contents
1 Introduction
1.1 Optimization
1.2 Definition of an Optimization Problem
1.3 Local and Global Optimal Solutions
1.4 Methods for Global Optimization
1.5 Nature Inspired Computing Techniques
1.6 The No Free Lunch Theorem
1.7 Harmony Search Algorithm
1.7.1 Harmony Search variants based on handling of parameter
1.7.1.1 Improved Harmony Search
1.7.1.2 Global Best Harmony Search
1.7.1.3 Adaptive Harmony Search algorithm
1.7.1.4 Self-adaptive Harmony Search
1.7.1.5 Self-adaptive Global Best Harmony Search
1.7.1.6 Other variants of Harmony Search based on handling of parameters
1.7.2 Variants based on hybridization of HS with other metaheuristic algorithms
1.7.3 Applications of Harmony Search Algorithm
1.8 Motivation and Objectives of the Thesis
1.9 Organization of the Thesis
2 Developments in Harmony Search Algorithm
2.1 Harmony Search variants based on handling of parameter
2.1.0.1 Improved Harmony Search
2.1.0.2 Global Best Harmony Search
2.1.0.3 Adaptive Harmony Search algorithm
2.1.0.4 Self-adaptive Harmony Search
2.1.0.5 Self-adaptive Global Best Harmony Search
2.1.0.6 Intelligent Tuned Harmony Search Algorithm
2.1.0.7 Improved Global-best Harmony Search
2.1.1 Other variants of Harmony Search based on handling of parameters
2.2 Variants based on hybridization of HS with other metaheuristic algorithms
3 Applications of Harmony Search Algorithm
3.1 Overview of Harmony Search Algorithm applications
4 A Hybrid Harmony Search and Simulated Annealing Algorithm for Continuous Optimization
4.1 Introduction
4.2 Simulated Annealing
4.3 Proposed Hybrid Harmony Search and Simulated Annealing (HS-SA) algorithm
4.4 Numerical Experiments on CEC 2014 benchmark suite
4.4.1 IEEE CEC 2014 Benchmark suite
4.4.2 Analysis of results
4.4.2.1 Convergence Behaviour
4.4.3 Wilcoxon rank test analysis
4.4.3.1 Algorithm Complexity
4.5 Conclusion
Objectives and Topics
The primary research objective is to develop efficient Harmony Search-based algorithms and evaluate their performance on standard benchmarks and real-world engineering problems to achieve a better balance between diversification and intensification during the optimization search process.
- Design of reliable Harmony Search variants including Two Phase and Shrinking Memory Harmony Search.
- Hybridization of Harmony Search with Simulated Annealing and Hill Climbing operators.
- Validation on complex IEEE CEC 2014 benchmark suites and multimodal functions.
- Application of the proposed techniques to real-world problems like Camera Calibration, Sudoku puzzle solving, and industrial engineering tasks.
Excerpt from the book
1.7 Harmony Search Algorithm
Harmony Search (HS) (Geem et al., 2001)is a musician’s behavior inspired evolutionary algorithm developed in 2001, though it is a relatively new meta heuristic algorithm, its effectiveness and advantages have been demonstrated in various applications.
Weyland (Weyland, 2012) raised an issue regarding the novelty of Harmony Search algorithm by declaring it a special case of (μ+1)−ES, however the pitch adjustment operator used in HS is entirely different than the mutation operator used in ES. Further HS utilizes the pitch adjustment operator (local search) probabilistically in contrast to ES’s mutation operator and thus the two can’t be considered same. Ample evidence has been provided in (Saka et al., 2016) to show HS is not a special case of (μ + 1)− ES even though superficially they seem to be identical.
In order to explain the Harmony Search in detail, let us first idealize the improvisation process by a skilled musician. When a musician is improvising there are three possible choices:
1. Play any piece of music exactly from his memory.
2. Play something similar to a known piece.
3. Compose new or random notes.
Summary of Chapters
1 Introduction: This chapter defines optimization problems, reviews existing literature on Nature Inspired Algorithms, and outlines the motivation and structure of the thesis.
2 Developments in Harmony Search Algorithm: This chapter reviews various modifications of the Harmony Search algorithm, focusing on dynamic parameter handling and hybrid variations found in the literature.
3 Applications of Harmony Search Algorithm: This chapter provides an overview of various fields where Harmony Search is successfully applied, including puzzle solving like Sudoku and complex optimization tasks.
4 A Hybrid Harmony Search and Simulated Annealing Algorithm for Continuous Optimization: This chapter proposes a novel hybrid HS-SA algorithm, detailing its mathematical formulation and evaluating its performance on the IEEE CEC 2014 benchmark suite.
Keywords
Harmony Search, Optimization, Metaheuristic Algorithms, Simulated Annealing, Global Optimization, Engineering Problems, Convergence, Hybridization, Benchmark Functions, Multimodal Functions, Computational Intelligence, Parameter Tuning, Exploration, Exploitation.
Frequently Asked Questions
What is the core focus of this research?
The research focuses on the theory and applications of the Harmony Search (HS) algorithm, specifically aiming to enhance its balance between exploration and exploitation through hybrid metaheuristic approaches.
What are the primary themes discussed?
The main themes include optimization techniques, nature-inspired computing, algorithm hybridization, and their specific application to continuous and combinatorial optimization problems.
What is the primary objective of the work?
The work aims to design more efficient and reliable Harmony Search-based algorithms and validate them against benchmark functions and real-world engineering challenges.
Which scientific methods are employed?
The study utilizes evolutionary computing, metaheuristic hybridization (specifically with Simulated Annealing and Hill Climbing), and statistical analysis via Wilcoxon rank-sum tests to validate performance.
What does the main body cover?
The main body covers a comprehensive review of existing Harmony Search variants, the proposal of a hybrid HS-SA algorithm, and detailed numerical experiments on IEEE CEC 2014 benchmark functions.
Which keywords characterize this thesis?
Key terms include Harmony Search, Optimization, Metaheuristics, Hybridization, Simulated Annealing, and various application-specific terms like camera calibration and truss structure optimization.
What is the benefit of the proposed HS-SA hybrid algorithm?
The HS-SA hybrid algorithm combines the exploitation capabilities of Harmony Search with the exploration benefits of Simulated Annealing, allowing the algorithm to escape local optima more effectively.
How is the algorithm's complexity addressed?
The time complexity is evaluated according to IEEE CEC 2014 standards, comparing the computing time and resource usage of HS-SA against standard Harmony Search and Simulated Annealing.
- Citar trabajo
- Assif Assad (Autor), 2018, Harmony Search Algorithm. Theory and Applications, Múnich, GRIN Verlag, https://www.grin.com/document/461697
