Catálogo de publicaciones - libros

This chapter introduces a relatively new meta-heuristic for combinatorial optimization, the ant colony. The ant colony algorithm is a multiple solution global optimizer that iterates to find optimal or near optimal solutions. Like its siblings genetic algorithms and simulated annealing, it is inspired by observation of natural systems, in this case, the behavior of ants in foraging for food. Since there are many difficult combinatorial problems in the design of reliable systems, applying new meta-heuristics to this field makes sense. The ant colony approach with its flexibility and exploitation of solution structure is a promising alternative to exact methods, rules of thumb and other meta-heuristics.

Optimization of reliability of complex systems is an extremely important issue in the field of reliability engineering. Over the past three decades, reliability optimization problems have been formulated as non-linear programming problems within either single-objective or multi-objective environment. Tillman .. (1980) provides an excellent overview of a variety of optimization techniques applied to solve these problems. However, he reviewed the application of only derivative-based optimization techniques, as metaheuristics were not applied to the reliability optimization problems by that time.

Telecommunication networks, oil platforms, chemical plants and airplanes consist of a great number of subsystems and components that are all subject to failure. Reliability theory studies the failure behavior of such systems in relation to the failure behavior of their components, which is often easier to analyze. However, even for the most basic reliability models, the overall reliability of the system can be difficult to compute. In this chapter we give an introduction to modern Monte Carlo methods for fast and accurate reliability estimation. We focus in particular on Monte Carlo techniques for network reliability estimation, and network design.

Plenty of optimization meta-heuristics have been designed for various purposes in optimization. They have also been extensively implemented in reliability engineering. For example, Genetic Algorithm (Coit and Smith, 1996), Ant Colony Optimization (Liang and Smith, 2004), Tabu Search (Kulturel-Konak, et al., 2003), Variable Neighbourhood Descent (Liang and Wu, 2005), Great Deluge Algorithm (Ravi, 2004), Immune Algorithm (Chen and You, 2005) and their combinations (hybrid optimization techniques) exhibited effectiveness in solving various reliability optimization problems.

The aim of this chapter is to illustrate the computational benefits in network reliability assessment which results from combining the modeling power of Cellular Automata [23] and Monte Carlo sampling and simulation [10, 19].

The reliability assessment of a system requires knowledge of how the system can fail, failure consequences and modeling, as well as selection of the evaluation technique [4].

Today’s engineering systems are sophisticated in design and powerful in function. Examples of such systems include airplanes, space shuttles, telecommunication networks, robots, and manufacturing facilities. Critical measures of performance of these systems include reliability, cost, and weight. Optimal system design aims to optimize such performance measures.

Computer-based artificial systems have been widely applied in nearly every field of human activities. Whenever people rely heavily on some product/technique, they want to make sure that it is reliable. However, computer systems are not as reliable as expected, and software has always been a major cause of the problems. With the increasing reliability of hardware and growing complexity of software, the software reliability is a rising concern for both developer and users. Software reliability engineering (SRE) has attracted a lot of interests and research in the software community and software reliability modeling is one major part of SRE research.

The lifecycle of a product is generally associated with two key players viz., a producer and an end-user. Although producers and end-users depend on each other, they have their own priorities. For a producer, the characteristics of a population of units are more important than the individual units, whereas for an end-user the opposite is true. The government makes laws for the population of a country, but a parent may be more concerned about its impact on the future of their individual children. A car manufacturer targets consistency in fuel efficiency for a population of cars, whereas a car owner has concerns about the fuel efficiency of his/her car. Similar differences in concerns also apply to tennis racquet manufacturer versus a tennis player or a cutting tool manufacturer versus cutting tool user.

A lot of methods and models in classical reliability theory assume that all probabilities are precise, that is, that every probability involved is perfectly determinable. Moreover, it is usually assumed that there exists some complete probabilistic information about the system and component reliability behavior.