By Topic

Genetic Algorithms for Control Systems Engineering, IEE Colloquium on

Date 28 May 1993

Filter Results

Displaying Results 1 - 11 of 11
  • Designing embedded parallel systems with parallel genetic algorithms

    Page(s): 7/1 - 7/2
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (132 KB)  

    Generic parallel genetic algorithms are developed with reference to the example of the real-time path planning problem for mobile robots. Most robot motion planners are used off-line: the planner is invoked with a model of the environment, it produces a path which is passed to the robot controller which in turn executes it. In general, the time necessary to achieve this loop is not short enough to allow the robot to move in a dynamic environment (moving obstacles). The goal is to try to reduce this time in order to be able to deal with real time path planning in dynamic environments. The authors use a method, called `Ariadne's CLEW algorithm', to build a global path planner based on the combination of two parallel genetic algorithms: an EXPLORE algorithm and a SEARCH algorithm. The purpose of the EXPLORE algorithm is to collect information about the environment with an increasingly fine resolution by placing landmarks in the searched space. The goal of the SEARCH algorithm is to opportunistically check if the target can be reached from any given placed landmark View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optimal population size for genetic algorithms: an investigation

    Page(s): 2/1 - 2/4
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (128 KB)  

    The performance of genetic algorithms (GAs) is affected by the parameters that are employed. In particular, the population size affects the performance and efficiency of GA-based systems. Grefenstette (1986) claimed that a population size between 60-110 is optimal for the convergence of GA-based systems to optimal solution. This paper presents studies that do not support this claim. GAPOLE, a GA-based program, is used to build self-learning self-adaptive self-optimising controllers for a dynamic multi-output unstable system using different population sizes. It is argued that population size may need to be tuned from one application to the other View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multiobjective genetic algorithms

    Page(s): 6/1 - 6/5
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (244 KB)  

    Multiobjective genetic algorithms (MOGAs) are introduced as a modification of the standard genetic algorithm at the selection level. Rank-based fitness assignment and the implementation of sharing in the objective value domain are two of the important aspects of this class of algorithms. The ability of the decision maker (DM) to progressively articulate its preferences while learning about the problem under consideration is one of their most attractive features. Illustrative results of how the DM can interact with the genetic algorithm are presented. They also show the ability of the MOGA to uniformly sample regions of the trade-off surface View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Genetic algorithms in control systems engineering: a brief introduction

    Page(s): 1/1 - 1/5
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (296 KB)  

    Genetic algorithms (GA) are adaptive search techniques, based on the principles of natural genetics and natural selection, which, in control systems engineering, can be used as an optimization tool or as the basis of more general adaptive systems. Following an introduction to the simple GA, important characteristics of GA are identified and control applications are described View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Systems identification with genetic algorithms

    Page(s): 3/1 - 3/3
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (116 KB)  

    Genetic algorithms are applied to the identification of black-box systems and partially known systems. The approach is best suited to the partially known systems (PKS) problem; in contrast to least-squares-based algorithms for identification of linear black-box systems, corresponding algorithms for identification of partially known systems are in the early stages of development. The best known algorithms for PKS identification suffer from local minima problems. It is shown that the genetic search and optimisation approach overcomes the local minima problem. Further, the approach is applicable immediately to multiparameter PKS identification problems without modification. This paper outlines a framework for black-box and PKS identification with genetic algorithms View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Genetic-based agents for control of distributed systems

    Page(s): 10/1 - 10/4
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (232 KB)  

    Management and control of distributed systems are hard tasks for conventional control techniques. Distributed computer control systems (DCCS) and data communication technology have helped to alleviate some of the problems of distributed systems. However, there are some problems that are difficult to solve using conventional methods. Artificial intelligence (AI) techniques have been proposed as solutions to many of the problems inherent in distributed systems. These solutions sometimes prove too complex to use in real systems, and a simpler adaptive system may be needed. This paper discusses adaptive systems from the genetic-based machine learning paradigm, and how they can be integrated with distributed artificial intelligence techniques for the control of distributed systems View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Classifier systems for control

    Page(s): 8/1 - 8/3
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (136 KB)  

    Classifier systems lie midway between neural networks and symbolic processing systems and potentially combine the benefits of both. They are parallel message-passing rule-based systems which use genetic algorithms to discover new rules as well as providing for reinforcement learning and programming. It has been proposed that a suitable application of genetic algorithms is to evolve robots. A most suitable way to use genetic algorithms to evolve the control systems for robots is within the framework provided by classifier systems. At a SERC workshop on learning systems a number of groups presented successful applications of the genetic algorithm to control problems. However, one cannot evolve complex systems with a simple genetic algorithm nor is it wise or safe to start from scratch in real applications where programmed knowledge can provide constraints for the genetic algorithm to work within. If the genetic algorithm is to be used to evolve control systems for industrial or commercial applications one of the best ways to do this is within the framework of classifier systems View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • IEE Colloquium on `Genetic Algorithms for Control Systems Engineering' (Digest No. 1993/130)

    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (12 KB)  

    The following topics were dealt with: algorithm optimal population; systems identification; robot trajectory planning and collision avoidance; multiobjective genetic algorithms; parallel genetic algorithms; classifier systems; distributed multivariable fuzzy control; and AI in distributed control View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Genetic-based minimum-time trajectory planning of articulated manipulators with torque constraints

    Page(s): 4/1 - 4/3
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (184 KB)  

    This paper is concerned with introducing a genetic-based algorithm for the minimum-time trajectory planning of articulated robotic manipulators. The planning procedure is performed in the configuration space and respects all physical constraints imposed on the manipulator design, including the limits on the torque values applied to the motor of each joint of the arm; consequently, the complete nonlinear dynamic robot model is incorporated in the formulation View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A distributed genetic algorithm for multivariable fuzzy control

    Page(s): 9/1 - 9/3
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (156 KB)  

    The traditional approach to multiple parameter optimization in genetic algorithm (GA) practice is to combine the coding of the parameters into a single compound bit-string; the so-called concatenated binary mapping. This approach has some shortcomings; the GA is a competition-based technique that has a natural tendency to evolve one winner which in complex problems yields a solution that is better on some parameters than the others. An extension to the simple GA, called vector evaluated genetic algorithm (VEGA), has been used in multiobjective optimization where one is not interested in a single solution, but a family of optimal solutions. In VEGA each member of the population is evaluated and assigned a weighted fitness value dependent on how it relates to each objective criteria. The reproduction plan then develops groupings within the populations for each of the objectives to be optimized, ensuring that the improvement of one objective does not adversely affect the others. This, however, requires large population sizes and can be quite inefficient. In cases where the complex task is divisible into simpler optimization problems, a better solution set may be obtained using parallel genetic algorithms to search for the optimal solution to each sub-problem View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Generation of collision-free paths, a genetic approach

    Page(s): 5/1 - 5/6
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (232 KB)  

    This paper deals with the generation of minimum distance paths for a robot manipulator operating in an environment cluttered with obstacles. These paths are optimised using a genetic approach. The genetic algorithm objective function is formulated in an obstacle avoidance problem context. Evaluation of the genetic algorithm parameters and their behaviour is undertaken in order to determine the most suitable values for this application. Finally cases involving first a mobile robot and then a two-link revolute manipulator are developed View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.