Swarm Intelligence - Volume 3: Applications
Swarm Intelligence (SI) is one of the most important and challenging paradigms under the umbrella of computational intelligence. It focuses on the research of collective behaviours of a swarm in nature and/or social phenomenon to solve complicated and difficult problems which cannot be handled by traditional approaches. Thousands of papers are published each year presenting new algorithms, new improvements and numerous real world applications. This makes it hard for researchers and students to share their ideas with other colleagues; follow up the works from other researchers with common interests; and to follow new developments and innovative approaches. This complete and timely collection fills this gap by presenting the latest research systematically and thoroughly to provide readers with a full view of the field of swarm. Students will learn the principles and theories of typical swarm intelligence algorithms; scholars will be inspired with promising research directions; and practitioners will find suitable methods for their applications of interest along with useful instructions. Volume 3 includes 27 chapters presenting real-world applications of swarm intelligence algorithms and related evolutionary algorithms. The companion volume 1 covers principles of swarm intelligence and volume 2 covers new algorithms and innovative methods. With contributions from an international selection of leading researchers, Swarm Intelligence is essential reading for engineers, researchers, professionals and practitioners with interests in swarm intelligence.
Inspec keywords: optimisation; evolutionary computation; swarm intelligence
Other keywords: finite-element model; evolutionary algorithms; swarm intelligence; Particle swarm optimization
Subjects: Expert systems and other AI software and techniques; Artificial intelligence (theory); Optimisation techniques
- Book DOI: 10.1049/PBCE119H
- Chapter DOI: 10.1049/PBCE119H
- ISBN: 9781785616310
- e-ISBN: 9781785616327
- Page count: 880
- Format: PDF
-
Front Matter
- + Show details - Hide details
-
p.
(1)
-
1 Prototype generation based on MOPSO
- + Show details - Hide details
-
p.
1
–29
(29)
When classifying a test instance, the nearest neighbor classifier will consume lots of time because it needs to search the whole training set for the instance's nearest neighbors. Prototype generation is a widely used approach to improve its time efficiency which generates a small set of prototypes to classify a test instance instead of using the whole training set. This paper applies the particle swarm optimization (PSO) to prototype generation and presents two novel methods to improve the classifier's performance. A fitness function named error rank is proposed to enhance the nearest neighbor classifier's generalization ability. In order to keep the classifier from overfitting the training set, this paper proposes the multiobjective optimization strategy which divides the whole training set into several subsets and regards the performance criterion on each subset as an objective function of the multiobjective PSO. The multiobjective optimization strategy pursues the performance over multiple subsets simultaneously, resulting in better generalization ability. Experimental results over 31 UCI datasets and 59 additional datasets show that the proposed algorithm achieves state-of-the-art performance.
-
2 Image reconstruction algorithms for electrical impedance tomography based on swarm intelligence
- + Show details - Hide details
-
p.
31
–55
(25)
Swarm intelligence algorithms have a wide range of application possibilities such as in the fields of engineering, economics, biology and medicine. One technique that emerges from the use of such algorithms is the electrical impedance tomography (EIT), which consists of a noninvasive and free of ionizing radiation imaging technique with applications in industry, geophysics and medicine. The technique is based on the application of a pattern of alternating low amplitude and high-frequency electric current through electrodes arranged around the surface of the section of the object to be imaged and in the consequent analysis of the electric potential measured by others electrodes. EIT technique consists of the solution to the direct and inverse problems. The direct problem consists in defining the electrical potentials within the object section and the potentials measured in its boundary by knowing the internal conductivity distribution of the object and the current excitation pattern, the relation of which is given by Laplace's equation. However, the estimation of the conductivity and electrical permittivity distribution of the interior of the body section from the measurements of the excitation response is mathematically an inverse, nonlinear and ill-posed problem. These characteristics make their solution quite dependent on the reconstruction and regularization algorithm and can be obtained through noniterative and interactive methods. The proposal of this chapter is to present the development of a low-cost hardware-software electrical impedance tomography system employing a project-partitioning strategy, developed for the acquisition and conditioning of data to preprocess and transfer the electrical potentials from the measured boundary of the object to a computer, performing the image reconstruction with swarm-based algorithms, namely, particle swarm optimization (PSO), artificial bee colony and fish school search.
-
3 A semisupervised fuzzy GrowCut algorithm for segmenting masses of regions of interest of mammography images
- + Show details - Hide details
-
p.
57
–84
(28)
According to the World Health Organization, breast cancer is the most common cancer in women worldwide, becoming one of the most fatal types of cancer. Mammography image analysis is still the most effective imaging technology for breast cancer diagnosis, which is based on texture and shape analysis of mammary lesions. The GrowCut algorithm is a general-purpose segmentation method based on cellular automata, able to perform relatively accurate segmentation through the adequate selection of internal and external seed points. This chapter shows an adaptive semisupervised version of the GrowCut algorithm, based on the modification of the automaton evolution rule by adding a Gaussian fuzzy membership function in order to model nondefined borders. In this proposal, manual selection of seed points of the suspicious lesion is changed by a semiautomatic stage, where just the internal points are selected by using a differential evolution algorithm. We evaluated the proposal using 59 lesion images obtained from MiniMIAS database. The results were compared with the semisupervised state-of-the-art approaches bidimensional empirical mode decomposition, breast mass contour segmentation, wavelet analysis, topographic approach, and marker-controlled watershed (MCW). The results show that fuzzy GrowCut achieves better results for circumscribed, spiculated lesions, and ill-defined lesions, considering the similarity between segmentation results and ground-truth images.
-
4 Multiobjective optimization of autonomous control for a biped robot
- + Show details - Hide details
-
p.
85
–107
(23)
In the present study, a multiobjective particle swarm optimization (PSO) is used to Pareto optimal design of controller for a biped robot walks in the coronal plane. Both nonlinearity of the dynamic equations and the tracking system cause an effective control that has to be used to address these problems.Thus, the proportional-derivative control is employed to control of the robot in the coronal plane, and the multiobjective algorithm is utilized to tune the heuristic parameters of the controller. The obtained Pareto front by the multiobjective PSO algorithm is compared with three prominent algorithms: modified NSGAII, sigma method, and MATLAB® Toolbox MOGA.
-
5 Swarm intelligence based MIMO detection techniques in wireless systems
- + Show details - Hide details
-
p.
109
–137
(29)
These nature-inspired techniques testified to be efficient maximum likelihood (ML) function optimizers. Their simple and less complex architectures make them worthy for non-deterministic polynomial (NP)-hard multi-input multi-output (MIMO) detection problem. Genetic algorithm, particle swarm optimization and ant colony optimization methods approach near optimal performance with significantly reduced computational complexity, particularly in the case of higher constellation systems and alphabet sizes with multiple transmitting antennas as compared to traditional ML detector that is computationally expensive and nonpractical to utilize. Swarm intelligence (SI) based mechanisms show their efficacy for solving MIMO detection problem as well as a promise for these heuristic algorithms to be applied in complex modulation mechanisms. One of the main contributions of the work in this chapter is to prove that SI is a useful optimization technique for classical communications issues for which these approaches were not considered very effective in the past.
-
6 Swarm intelligence in logistics and production planning
- + Show details - Hide details
-
p.
139
–166
(28)
In this chapter, major contributions of swarm intelligence in the fields of logistics and production planning are discussed. Starting with a general introduction to planning problems in these fields, we outline the limitations of traditional optimization approaches and the reasons for using methods from the field of swarm intelligence such as the NP-hardness of many important problems (Section 6.1). We discuss some general aspects of utilizing swarm algorithms which can be used for optimizations problems in logistics and productions, and introduce briefly some well-established and a few newer approaches in that field (Section 6.2). After that, the most important problem types such as lot-sizing problems, scheduling problems, and vehicle routing problems are discussed including modeling aspects and results from swarm intelligence applications (Section 6.3). As a result, we see that established approaches such as particle swarm optimization and ant colony optimization are well established in these areas including various variants and improvements, which were worked out for the specific problems under consideration including hybridizations of the algorithms with other techniques. We also discuss the current situation with respect to solving such problems in real life including the future potential of including swarm intelligence in commercial solutions (Section 6.4).
-
7 Swarm intelligence for object-based image analysis
- + Show details - Hide details
-
p.
167
–188
(22)
Object-based image analysis (OBIA) approaches are often superior to the pixel-based classifications at very high resolution (VHR) remotely sensed images. Due to the similar spectral signatures of land-cover classes especially in urban areas, spatial information must be exploited to produce improved classification maps at finer resolutions. Segmentation and rule-based classification are the main two steps of the powerful OBIA approach which is widespread in pattern recognition and classification applications. Selecting the best values for segmentation parameters has an important effect on the segmentation results. Once the image objects are derived, topological relations between them, statistical summaries of spectral and textural features and shape features can all be employed in the rule-based classification. Optimum feature selection also has an essential role for rule set generation. Thus, optimal parameter/feature selection may be an important process in both steps of the OBIA approach. Among other optimization techniques, metaheuristic optimization algorithms such as swarm-intelligence-based methods are very capable of solving feature selection problems. So, they can be used inboth steps of OBIA approaches. In segmentation, the capabilities of swarm intelligence may optimize the parameters. Moreover, ant colony optimization (ACO) and particle swarm optimization (PSO) are successfully utilized for optimum feature selection in forming rule-based classification. In the first two sections of this chapter, the necessity of performing OBIA in object recognition based on VHR images is explained. Then, the basis of optimum feature selection and optimization algorithms is mentioned. After a comprehensive review on the concepts of ACO, PSO and firefly algorithm (FA) as the powerful swarm intelligence algorithms, the capabilities of these algorithms are investigated in the field of optimum feature selection. Finally, the experimental results of performing the FA and PSO for optimum feature selection are investigated and generalized for improving the capabilities of the OBIA approaches.
-
8 Evolutionary multiobjective optimization for multilabel learning
- + Show details - Hide details
-
p.
189
–218
(30)
Machine learning is traditionally considered as a single objective optimization problem. That is, a learner iteratively improves performances through optimizing one single objective function. However, many machine-learning tasks need to simultaneously optimize multiple objective functions. For example, ensemble learning need to find a set of accurate but diverse base learners, and it is desirable for the feature selection task to extract a small number of representative features. As a consequence, there is a surge of multiobjective machine learning in recent years. That is, a machine-learning task is considered as the multiobjective optimization problem (MOP). Moreover, evolutionary algorithms are usually employed to optimize it, since evolutionary algorithms have shown their superiority for MOPs. In this chapter, we will introduce our recent work in multilabel learning with multiobjective optimization. As an important supervised learning task, multilabel learning refers to the task of predicting potentially multiple labels for a given instance. Conventional multilabel learning approaches focus on a single objective setting, and there is a basic assumption that the optimization over one single objective can improve the overall performance of multilabel learning. However, in many real applications, an optimal multilabel learner may need to consider the trade-offs among multiple objectives. In this chapter, we will present two works in evolutionary multiobjective optimization for multilabel learning. In the first work, we directly optimize multiobjective functions based the multiobjective optimization framework; in the second work, we generate a set of accurate but diverse multilabel learners with the ensemble learning framework.
-
9 Image segmentation by flocking-like particle dynamics
- + Show details - Hide details
-
p.
219
–244
(26)
In this chapter, we present a segmentation technique based on patterns that emerge from a system of moving particles. It is inspired from the flocking formation that can be frequently observed in nature, e.g. in large groups of birds and fish, which display a complex, coordinated motion without the guidance of any leader. First, we down-sample the image by dividing it into superpixels, each of which will be represented by a moving particle. After assigning random directions of motion for every particle, the dynamical evolution starts; after some iterations, it is expected that the new particle arrangement reflects useful image features, i.e. the segments we search for. We also present a comprehensive parameter analysis by objectively measuring our results compared to human-annotated ground-truth images. A comparison to some other segmentation algorithms is also performed, what enables us to discuss some pros and cons of our flocking-like approach. Rather than just present a new segmentation algorithm, our intention is to bring ideas that may motivate new studies and applications of self-organising dynamical systems to solve complex problems that involve large datasets.
-
10 Swarm intelligence for controller tuning and control of fractional systems
- + Show details - Hide details
-
p.
245
–281
(37)
In this chapter, particle swarm optimization, ant colony optimization, artificial bee colony optimization and cuckoo search optimization are examined in detail. Differences, advantages and disadvantages of these algorithms are emphasized clearly. Performances of the swarm algorithms in terms of computing time, computing complexity and accuracy and convergence behavior are compared each other. As application area, fractional control systems from new attractive topics of control are chosen. In this chapter, the fractional order proportional integral derivative controllers are tuned with the swarm algorithms using objective functions such as integral of absolute error, integral of the squared error, the integral of time multiplied by the absolute error and integral of time multiplied by the squared error. The simulation results can be used to determine which swarm algorithms yield better search performance in the multiobjective and high-dimensional nonlinear constrained optimization problems such as the fractional order control systems.
-
11 PSO-based implementation of smart antennas for secure localisation
- + Show details - Hide details
-
p.
283
–326
(44)
This chapter presents a flexible approach towards the design and implementation of smart antennas using particle swarm optimisation technique. Effectiveness of this evolutionary algorithm has been justified here with its two-fold applications in solving complex antenna optimisation problems. Initially, the design optimisation of a low-profile rectangular microstrip patch antenna on IE3D software has been demonstrated. Finally, adaptive beamforming in smart antennas to produce beam pattern with pre-specified side-lobes level and nulls at various interference directions has been described. The patch antenna was fabricated on a printed circuit board material of RT/duroid® 5880 laminates with dielectric constant of 2.4 and substrate thickness of 1.5875 mm. The important antenna parameters such as gain, voltage standing wave ratio, directivity, return loss, etc. were measured with vector network analyser and antenna test bench. The adaptive beamforming circuit was realised with finite state machine modelling on Virtex4 field programmable gate array board. Its performance was also verified with hardware level fixed point simulation on beamforming accuracy and computational overheads underboth of additive white Gaussian noise and Rayleigh fading channel conditions. This antenna system is suitable for secure localisation applications under harsh radio environment in wireless sensor networks (WSNs).
-
12 Evolutionary computation for NLP tasks
- + Show details - Hide details
-
p.
327
–367
(41)
Natural language processing (NPL) is an important research field that deals with very different problems (tasks) concerning how to interpret, generate and process natural language. Different approaches have been proposed to tackle these problems. More recently, a significant number of works that use evolutionary computation to solve some of them were presented. Among these, we can find attempts to solve the problem of word segmentation, part-of-speech tagging, syntactic sentence analysis and grammar generation. Despite the good results obtained by these approaches, these techniques are still not widely used by the community of researchers working in the area of NPL. With this chapter, we aim to contribute to the dissemination of these relatively recent global optimisation techniques as valid alternatives to the classic approaches normally used to tackle these problems. To achieve this, we begin by making a description of these algorithms, sufficiently exhaustive, in our opinion, to understand their fundamental aspects. Next, we present, in detail, the most representative works found in the literature that apply evolutionary computation-based techniques to the tasks mentioned above.
-
13 Particle swarm optimisation for antenna element design
- + Show details - Hide details
-
p.
369
–402
(34)
Particle swarm optimisation is one of the many optimisation techniques available and is an evolutionary computation technique inspired by social behaviour of bird flocking or fish schooling. This chapter presents a brief introduction to particle swarm optimisation and, in particular, real number, binary and hybrid particle swarm optimisations along with their application to antenna element design. This chapter also provides examples of antenna designs using particle swarm optimisation and comparisons with practical counterparts.
-
14 Swarm intelligence for data mining classification tasks: an experimental study using medical decision problems
- + Show details - Hide details
-
p.
403
–428
(26)
In the data-mining field, classifier learning techniques focus on extracting the knowledge from a problem using a set of labeled objects in order to predict the class for any other new object. Even though there exist several approaches to build these classifiers, some learning paradigms guide the creation of classifiers through an optimization process of a measure computed over the data. In this context, function optimization methods based on swarm intelligence can make the most of the interactions between individuals and be employed as a tool to explore a variety of potential solutions until reaching the convergence to a final and accurate solution for the classification problem. The present chapter, following the philosophy of this book, is aimed at guiding the reader through the whole process, from concepts to applications, when we refer to swarm intelligence applied to classification tasks. Thus, firstly, it presents a detailed explanation on how swarm intelligence algorithms can be used in classifier-building processes. In order to accomplish this, two of the most well-known classification techniques within this group of methods, which belong to particle swarm and ant colony optimizations, are analyzed. Second, the application of such methods for medical data classification is studied considering several real-world datasets. The results obtained show that swarm intelligence methods can play an important role for data analysis in medical applications, providing good performance results and models characterized by a high interpretability.
-
15 Towards spiking neural systems synthesis
- + Show details - Hide details
-
p.
429
–461
(33)
In this chapter, authors present the novel employment of swarm intelligence in the constraint design of spiking neurons and synapses for advanced and technology agnostic sensor signal conditioning and conversion. Manual design activities, leading to proof-of-principle circuits and a first chip, are used as specifications and baseline for the optimized designs pursued in our correspondingly extended ABSYNTH design environment.
-
16 Particle swarm optimization based memetic algorithms framework for scheduling of central planned and distributed flowshops
- + Show details - Hide details
-
p.
463
–494
(32)
In this chapter, we provide a panorama of the PSO-based memetic algorithm (MA) for traditional permutation flowshop scheduling problem (PFSP) and its several variants. In the proposed algorithm, the global exploration ability of PSO and the local refinement ability of simulated annealing (SA) are delicately integrated and balanced. Some specific techniques related to the nature of PFSP are introduced to further improve the effectiveness of PSO-based MA. The key features in the proposed algorithm are detailed as follows. First, to apply PSO in solving combinatorial optimization problems such as PFSP, we rely on the ranked-order value (ROV) rule that uses random key representation to transform the continuous position information to scheduling permutations. Second, NEH and NEH-based constructive heuristics are introduced to guarantee a proportion of initial particles to be of good qualities. Third, to avoid the premature convergence problem of PSO, an adaptive SA-based local search is proposed to strengthen the exploitation in an efficient way. Forth, for the variation of PFSP that considers distributed processing factories, single assembly factory, and no-wait constraint (DAPFSP-NW), we include an extra encoding layer to represent the factory dispatch; thus, the proposed SA-based MA can still be applied. Moreover, the corresponding heuristic-based initialization and the neighborhoods adopted for local search are redefined. Last but not the least, for the variation with stochastic processing time and assembling time, the technique of hypothesis test (HT) is integrated into PSO-based MA; therefore, the solutions generated in each iteration can be effectively compared in a statistical way. Our experimental results strongly indicate the superiority of the PSO-based MA for solving PFSP, DAPFSP-NW, and stochastic DAPFSPNW at the aspects of optimization quality and robustness.
-
17 Particle swarm optimization for antenna array synthesis, diagnosis and healing
- + Show details - Hide details
-
p.
495
–530
(36)
In modern wireless communication systems, the antenna array is an integral part to achieve the spatial diversity. The increased demand of these arrays, mostly in radars and space applications, has attracted a lot of researchers in Microwave Engineering in-general and Antenna Engineering in-particular, to put in their research efforts in this area. The outcome is, not only the innovative analysis techniques, but also a lot of flexibility in handling these arrays. The research focus is directed towards the development of self-healing arrays, where it can be possible to control the array placed in a space platform, find the faults, if any, in it, and automatically apply a suitable compensation for the fault. The present chapter focuses these aspects of antenna arrays as an application of particle swarm optimization (PSO). An approach with the use of PSO has been made by the authors for antenna array synthesis, fault diagnosis and a corresponding compensation. The chapter will start with a brief on antenna arrays, their analysis and synthesis techniques. Reason and suitability of PSO over other analytical and swarm intelligence based optimization techniques will be followed. Then the problem formulation, solution and results of the developed process using PSO will be discussed in length, which include the analysis of antenna arrays, fault finding and compensation procedure. The chapter will conclude with the overall analysis results and the future research directions.
-
18 Designing a fuzzy logic controller with particle swarm optimisation algorithm
- + Show details - Hide details
-
p.
531
–555
(25)
In this study, we designed a trajectory-tracking fuzzy logic controller (TTFLC) for the TRIGA Mark-II Training and Research Reactor, which is located at the Istanbul Technical University. The designed fuzzy logic controller (FLC) is based on the zeroorder Sugeno method. The parameters of the FLC membership functions and the action weights of the 15 rules in the rule base are optimised by using the particle swarm optimisation (PSO) algorithm. The objective of this study is to control the TRIGA Mark-II reactor using the designed PSO-tuning TTFLC in a simulator. We used a simulation code from the literature called `YAVCAN' for studying the non-linear behaviour of the core of the TRIGA Mark-II reactor. To select the best parameters of the PSO algorithm for this system, we conducted some experiments. After selecting the best PSO parameters, the algorithm was started a number of times to determine the optimal parameters of the designed FLC and the optimal parameters of the controller. After determining these parameters, the performance of the designed controller was tested for various initial and desired power levels and under conditions of disturbance. The simulation results showed that the proposed controller could control the reactor power successfully, and it could ensure that the reactor tracks the desired trajectory power within the acceptable error tolerance. Therefore, the PSO algorithm is suitable for finding the optimal parameters of the FLC.
-
19 Adding swarm intelligence for slope stability analysis
- + Show details - Hide details
-
p.
557
–580
(24)
Since the introduction of particle swarm optimization (PSO), there has been a dramatic proliferation of published work on the theory and applications of swarm intelligence. However, the use of PSO in slope stability analysis (landslides) has been limited due to factors such as the lack of domain knowledge on the researcher's side and the complexity inherent in the problem itself. To identify issues and provide clear examples for the analysis of slopes and landslides, this chapter will review the existing literature available on the application of PSO to slope stability analysis and discuss a framework to integrate swarm intelligence with the computation of the factor of safety (FS) of slopes. This novel method involves adding swarm intelligence to the existing STABL program (created by Purdue University for general solutions of 2D slopes), and it does not require access to the source codes. It is believed that this method is particularly suitable for complex problems such as slope stability analysis where domain knowledge is important and writing a replacement program with PSO capability from the ground up is too time-consuming and too vulnerable to mistakes. The potential of this approach is of great significance because it could possibly augment and extend the utility of many existing programs in many application fields. To demonstrate the usefulness of this method, we applied it to both theoretical and real-world slopes and landslides and found that they both had good results (but different convergence points). It proved that our approach was both feasible and efficient. We believe that the proposed method will not only enhance the original computer program so that it can solve more problems in its application domain, but it will also benefit the field of swarm intelligence by providing more optimization examples from a wider range of applications.
-
20 Software module clustering using particle swarm optimization
- + Show details - Hide details
-
p.
581
–601
(21)
Software module clustering problem (SMCP) is an important problem of software engineering field. The large-scale SMCPs are very difficult to solve by using the traditional deterministic optimization methods within a reasonable amount of time. The stochastic metaheuristic search optimization methods have been found to be an effective alternative to address the SMCPs in reasonable computation cost. Recently, particle swarm optimization (PSO) algorithm a metaheuristic search optimization method has gained wide attention toward research community and has been demonstrated as an effective and convenient algorithm to solve the various science and engineering problems. To the best of our knowledge, the applicability and usefulness of the PSO algorithm have not been studied by any researcher till date to address the SMCPs. In this paper, we present a module clustering approach for restructuring the software system using the PSO algorithm. To evaluate the proposed software module clustering approach, six real-world software systems are restructured and the obtained clustering solutions are compared with clustering solutions obtained with existing state-of-the-art software module clustering algorithms (i.e., genetic algorithm, hill climbing, and simulated annealing) in terms of modularization quality (MQ), coupling, and cohesion. The statistical analysis of the MQ, coupling, and cohesion results of the clustering solution provides sufficient evidence that the proposed approach is able to generate more effective clustering solution compared to the existing state-of-the-art software module clustering algorithms.
-
21 A swarm intelligence approach to harness maximum techno-commercial benefits from smart power grids
- + Show details - Hide details
-
p.
603
–637
(35)
Modern electricity power market has evolved from conventional vertically integrated structure to the present deregulated form and eventually to the development of the smart power grids, over the past few decades. Different market players like generation, transmission and distribution companies with their individual and collective goals and constraints are now participating in real time, prompting the need for optimal allocation and utilization of the smart grid infrastructure and resources. The objective of the optimization thus received a paradigm shift from the traditional generation cost optimization, to optimal utilization of the available resources to deliver maximum benefit to all the power market participants and the so-called social welfare. Maximization of social welfare is a highly nonlinear optimization problem and generally requires application of an efficient stochastic optimization method with in-built ability of avoiding local optima. The swarm intelligence-based optimization algorithms, developed and presented in this book chapter, offer substantial improvement in the quality of solution to the problem over the conventional solution methods. The chapter presents real-time simulation and experiments on benchmark power system networks with the developed optimization algorithms. The results are found to be quite encouraging.
-
22 Fuzzy adaptive tuning of a particle swarm optimization algorithm for variable-strength combinatorial test suite generation
- + Show details - Hide details
-
p.
639
–662
(24)
Combinatorial interaction testing (CIT) is an important software testing technique that has seen lots of recent interest. It can reduce the number of test cases needed by considering interactions between combinations of input parameters. Empirical evidence shows that it effectively detects faults, in particular, for highly configurable software systems. In real-world software testing, the input variables may vary in how strongly they interact; variable strength CIT (VS-CIT) can exploit this for higher effectiveness. The generation of variable strength test suites is a non-deterministic polynomialtime (NP) hard computational problem. Research has shown that stochastic population-based algorithms such as particle swarm optimization (PSO) can be efficient compared to alternatives for VS-CIT problems. Nevertheless, they require detailed control for the exploitation and exploration trade-off to avoid premature convergence (i.e., being trapped in local optima) as well as to enhance the solution diversity. Here, we present a new variant of PSO based on Mamdani fuzzy inference system (FIS), to permit adaptive selection of its global and local search operations. We detail the design of this combined algorithm and evaluate it through experiments on multiple synthetic and benchmark problems. We conclude that fuzzy adaptive selection of global and local search operations is, at least, feasible as it performs only second-best to a discrete variant of PSO, called discrete PSO (DPSO). Concerning obtaining the best mean test suite size, the fuzzy adaptation even outperforms DPSO occasionally. We discuss the reasons behind this performance and outline relevant areas of future work.
-
23 Multiobjective swarm optimization for operation planning of electric power systems
- + Show details - Hide details
-
p.
663
–703
(41)
The determination of a safe and economic operating point for a power system lies in the definition of a set of controls (in this case, the optimal dispatch) that minimizes operation costs while ensuring system stability during each preestablished contingency. This is a problem that can be described as a case of multiobjective approach to the security constrained optimal power flow. This paper compares the performance of two multiobjective evolutive optimization techniques on obtaining this safe and economic operating point, according to static security point of view: the traditional nondominated sorting genetic algorithm (GA), based on GAs, and the multiobjective evolutionary particle swarm optimization (MOEPSO) developed in the scope of this chapter in order to exploit the performance gains observed with the hybrid metaheuristic evolutionary particle swarm optimization (EPSO) in single-objective optimization problems. The results of both multiobjective metaheuristics were parameterized so as to make possible the comparison with a monoobjective approach to the same problem. It was verified that both metaheuristics multiobjective presented improvement of performance after the implementation of a previous refinement of the initial solutions. It was also observed that the MOEPSO obtained the best performance among the other metaheuristics tested after the implementation of a methodology to obtain the cultural operators of the swarm based on the calculation of the Euclidean distance. The algorithms were implemented in MATLAB® and tested in a test-case that simulates the conditions of the Brazilian National Interconnected System (from the Portuguese Sistema Interligado Nacional).
-
24 Perturbed-attractor particle swarm optimization for image restoration
- + Show details - Hide details
-
p.
705
–724
(20)
This chapter addresses the problem of restoration of images degraded by blurring and noise. Image degradation is modeled as the convolution of a low-pass, two-dimensional point spread function with a reference or undegraded image. The image restoration problem is modeled as an optimization problem and a modified particle swarm optimization algorithm is developed. This approach is evaluated using image quality metrics that consider the quality of edges restored as well as the fidelity of the restored image with respect to the undegraded reference image. A novel-edge sharpness metric is also proposed. The new approach is shown to perform well in several image restoration examples.
-
25 Application of swarm intelligence algorithms to multi-objective distributed local area network topology design problem
- + Show details - Hide details
-
p.
725
–754
(30)
Artificial bee colony (ABC) optimization, ant colony optimization (ACO), and particle swarm optimization (PSO) are well-known swarm intelligence algorithms. They have been widely used for solving many real-life optimization problems in various domains. This chapter presents how these algorithms can be used in optimizing the distributed local area network topology design. The problem has been modelled as a constrained multi-objective optimization problem using goal programming. In addition to adapting the three algorithms for the problem, a hybrid ABC algorithm has also been proposed. Performance of the algorithms has been evaluated through a simulation study, and the results indicate that the hybrid ABC algorithm outperforms ACO, PSO and ABC algorithms.
-
26 Swarm intelligence algorithms for antenna design and wireless communications
- + Show details - Hide details
-
p.
755
–784
(30)
The purpose of this chapter is to briefly describe different swarm intelligence algorithms and present their application to antenna and wireless communications design problems. This chapter presents results from design cases that include linear antenna array synthesis for null suppression, thinned arrays, patch antenna, and a partial transmit sequence scheme for peak-to-average-power ratio (PAPR) reduction of orthogonal frequency division multiplexing signals.
-
27 Finite-element model updating using swarm intelligence algorithms
- + Show details - Hide details
-
p.
785
–813
(29)
In this chapter, several nature-inspired optimization algorithms are used to update finite-element models (FEMs) of structural systems. Usually, the numerical models of real mechanical structures, which are obtained by the FEM approach, give different results compared to the experimental measurements. The mismatch between numerical and experimental results is caused by the variability of the model parameters as well as the mathematical simplifications made during the modeling process. The procedure of correcting the numerical models is known as model updating where several model parameters are adjusted to minimize the error between the measurements and the numerical model. In this chapter, the model-updating procedure is defined as an optimization problem where several swarm intelligence algorithms: particle swarm optimization (PSO), ant colony optimization (ACO) and fish school search (FSS) algorithms are used to update the FEMs of two structural systems: A five degree of freedom (DOF) mass-spring system and an unsymmetrical H-shaped structure with real measurements. The results obtained in this study are compared with the results obtained by the genetic algorithm (GA). As a result, the updating procedures based on FSS, ACO and PSO algorithms give better results than the GA approach. Furthermore, the updating problem, in this chapter, is reformulated as a multiobjective (MO) problem, and a multiobjective PSO (MOPSO) algorithm was used to update the five DOF mass-spring system. The MOPSO algorithm shows promising result in model updating.
-
Back Matter
- + Show details - Hide details
-
p.
(1)