A Performance Comparison between Efficiency and Pheromone Approaches in Dynamic Manufacturing Scheduling

This chapter proposes an innovative coordination mechanism in manufacturing systems by pheromone approach in a multi-agent architecture environment. A pheromone-based coordination mechanism can reduce the communication among agents and decision-making complexity. The chapter focuses on job shop scheduling problem in cellular manufacturing systems. The principal aim is the evaluation of the performance of the proposed approaches compared with the approaches proposed in the literature (benchmark) in order to evidence the improvements. A simulation environment developed in ARENA® package was used to investigate the influence of several parameters on the manufacturing performance. The proposed approaches are tested in a dynamic environment; the simulation scenarios are characterized by the following parameters: inter-arrival, machine breakdowns and processing time efficiency. The simulation results highlighted that the performance of the proposed approaches are very competitive to the benchmark.

Diffusion and Emergence in Social Networks

Diffusion is the process by which new products and practices are invented and successfully introduced into society. Numerous studies on the diffusion of individual innovations have been conducted, many exhibiting common features such as the famous S-shaped diffusion curve. One basic question posed by innovation diffusion is why there is often a long lag time between an innovation’s first appearance and the time when a substantial number of people have adopted it. An extensive amount of theoretical and empirical literature has been devoted to this phenomenon and the mechanisms behind it. New ideas, products, and innovations often take time to diffuse, a fact that is often attributed to the heterogeneity of human populations. In this chapter, we provide an overview of the research examining how the structure of social networks impacts the diffusion process. The diffusion process enhances innovations via feedback of information about the innovation’s utility—which can be used to make future improvements—to many different users. This aspect of the diffusion process is similar to the micro-macro loop, which is an essential part of emergence. The aim of this research is to understand how the structure of social networks determines the dynamics of various types of emergent properties occurring within those networks. For emergence at the social level, patterns of social interactions are critical.

Synthesis of Analog Circuits by Genetic Algorithms and their Optimization by Particle Swarm Optimization

This chapter shows the application of particle swarm optimization (PSO) to size analog circuits which are synthesized by a genetic algorithm (GA) from nullor-based descriptions. First, a historical description of the development of automatic synthesis techniques to design analog circuits is presented. Then, the synthesis of analog circuits by applying a GA at the transistor level of abstraction is demonstrated. After that, the authors present the proposed multi-objective (MO) PSO algorithm which makes calls to the circuit simulator HSPICE to evaluate performances until optimal sizes of the transistors are found by using standard CMOS technology of 0.35µm of integrated circuits. Finally, the MO-PSO algorithm is compared with NSGA-II, and some open problems oriented to circuit synthesis and sizing are briefly discussed.

Evolutionary Based Adaptive User Interfaces in Complex Supervisory Tasks

In this chapter, the author proposes a novel idea based on evolutionary algorithm for adaptation of the user interface in complex supervisory tasks. Under the assumption that the user behavior is stationary and that the user has limited cognitive and motor abilities, the author has shown that a combination of genetic algorithm for constrained optimization and probabilistic modeling of the user may evolve the adaptive interface to the level of personalization. The non-parametric statistics has been employed to evaluate the feasibility of the ranking approach. The method proposed is flexible and easy to use in various problem domains. The author has tested the method with an automated user and a group of real users in an air traffic control environment. The automated user, implemented for initial tests, is built under the same assumptions as a real user. In the second step, the author has exploited the adaptive interface through a group of real users and collected subjective ratings using questionnaires. The author has shown that the proposed method can effectively improve human-computer interaction and his approach is pragmatically a valid design for the interface adaptation in complex environments.

Evolutionary Approaches and Their Applications to Distributed Systems

The ubiquitous presence of distributed systems has drastically changed the way the world interacts, and impacted not only the economics and governance but also the society at large. It is therefore important for the architecture and infrastructure within the distributed environment to be continuously renewed in order to cope with the rapid changes driven by the innovative technologies. However, many problems in distributed computing are either of dynamic nature, large scale, NP complete, or a combination of any of these. In most cases, exact solutions are hardly found. As a result, a number of intelligent nature-inspired algorithms have been used recently, as these algorithms are capable of achieving good quality solutions in reasonable computational time. Among all the nature-inspired algorithms, evolutionary algorithms are considerably the most extensively applied ones. This chapter presents a systematic review of evolutionary algorithms employed to solve various problems related to distributed systems. The review is aimed at providing an insight of evolutionary approaches, in particular genetic algorithms and genetic programming, in solving problems in five different areas of network optimization: network topology, routing, protocol synthesis, network security, and parameter settings and configuration. Some interesting applications from these areas will be discussed in detail with the use of illustrative examples.

A Lyapunov Theory-Based Neural Network Approach for Face Recognition

This chapter presents a new face recognition system comprising of feature extraction and the Lyapunov theory-based neural network. It first gives the definition of face recognition which can be broadly divided into (i) feature-based approaches, and (ii) holistic approaches. A general review of both approaches will be given in the chapter. Face features extraction techniques including Principal Component Analysis (PCA) and Fisher’s Linear Discriminant (FLD) are discussed. Multilayered neural network (MLNN) and Radial Basis Function neural network (RBF NN) will be reviewed. Two Lyapunov theory-based neural classifiers: (i) Lyapunov theory-based RBF NN, and (ii) Lyapunov theory-based MLNN classifiers are designed based on the Lyapunov stability theory. The design details will be discussed in the chapter. Experiments are performed on two benchmark databases, ORL and Yale. Comparisons with some of the existing conventional techniques are given. Simulation results have shown good performance for face recognition using the Lyapunov theory-based neural network systems.

Ant Colony Programming

Automatic programming is the method in which a computer program is constructed automatically based on the specification of goals which are to be realized. This chapter describes one of the methods for automatic function approximation (as a form of automatic programming) – ant colony programming (ACP). It is based on ant colony system (ACS) as a new method for solving approximation problems. While solving these problems by ACP two approaches are used: the expression approach and the program approach. Several improvements of this method are presented, including the elimination of introns, the use of a structure similar to the candidate list introduced in ACS, and parameter-tuning. The chapter first describes ACS and introduces the problem of symbolic regression. Then, ACP is defined. After that, improvements of ACP are presented. The main objective of the chapter is to give an overview of the published results of studies carried out on ACP, while at the same time present a new idea in the process of parameter-tuning.

Statistical Analysis of Computational Intelligence Algorithms on a Multi-Objective Filter Design Problem

This chapter presents the application of a comprehensive statistical analysis for both algorithmic performance comparison and optimal parameter estimation on a multi-objective digital signal processing problem. The problem of designing optimum digital finite impulse response (FIR) filters with the simultaneous approximation of the filter magnitude and phase is posed as a multi- objective optimization problem. Several computational-intelligence-based algorithms for solving this particular optimization problem are presented: genetic algorithms (GA), particle swarm optimization (PSO) and simulated annealing (SA) with multi-objective scalarization methods. Algorithms with Pareto sampling methods, namely non-dominated sorting genetic algorithm II (NSGA-II) and multi-objective simulated annealing (MOSA) are also applied as a way of dealing with multi-objective optimization. Instead of using a process of trial and error, a statistical exploratory analysis is used to estimate optimal parameters. A comprehensive statistical comparison of the applied algorithms is addressed, which indicates a particularly strong performance of NSGA-II and pure GA with weighting scalarization.

Efficient Training Algorithm for Neuro-Fuzzy Network and its Application to Nonlinear Sensor Characteristic Linearization

An ideal linear sensor is one for which input and output values are always proportional. Typical sensors are, in general, highly nonlinear or seldom sufficiently linear enough to be useful over a wide range or span of interest. Due to the requirement of tedious effort in designing sensor circuits with sufficient linearity for some applications, the word nonlinearity has acquired a pejorative connotation. Hence, a computationally intelligent tool for extending the linear range of an arbitrary sensor is proposed. The linearization technique is carried out by a very efficiently trained neuro-fuzzy hybrid network which compensates for the sensor’s nonlinear characteristic. The training algorithm is very efficient in the sense that it can bring the performance index of the network, such as the sum squared error (SSE), down to the desired error goal much faster than any first order training algorithm. Linearization of a negative temperature coefficient thermistor sensor with an exponentially decaying characteristic function is used as an application example, which demonstrates the efficacy of the procedure. The proposed linearization technique is also applicable for any nonlinear sensor (such as J-type thermocouple or pH sensor), whose output is a monotonically increasing/decreasing function.

A Review on Evolutionary Prototype Selection

Evolutionary algorithms have been successfully used in different data mining problems. Given that the prototype selection problem could be seen as a combinatorial problem, evolutionary algorithms have been used to solve it with promising results. This chapter presents an evolutionary data mining application known as evolutionary prototype selection. Various approaches have been proposed in the literature following two strategies on the use of evolutionary algorithms: general evolutionary models and models specific to prototype selection problem. In this chapter, the authors review the representative evolutionary prototype selection algorithms proposed, give their description and analyze their performance in terms of efficiency and effectiveness. They study their performance considering different sizes of the data sets, and analyze their behavior when the database scales up. The results are statistically contrasted in order to argue the benefits and drawbacks of each model.