Abdollah Homaifar

Simultaneous design of membership functions and rule sets for fuzzy controllers using genetic algorithms

This paper examines the applicability of genetic algorithms (GAs) in the simultaneous design of membership functions and rule sets for fuzzy logic controllers. Previous work using genetic algorithms has focused on the development of rule sets or high performance membership functions; however, the interdependence between these two components suggests a simultaneous design procedure would be a more appropriate methodology. When GAs have been used to develop both, it has been done serially, e.g., design the membership functions and then use them in the design of the rule set. This, however, means that the membership functions were optimized for the initial rule set and not the rule set designed subsequently. GAs are fully capable of creating complete fuzzy controllers given the equations of motion of the system, eliminating the need for human input in the design loop. This new method has been applied to two problems, a cart controller and a truck controller. Beyond the development of these controllers, we also examine the design of a robust controller for the cart problem and its ability to overcome faulty rules. >

Constrained Optimization Via Genetic Algorithms

This paper presents an application of genetic algorithms (GAs) to nonlinear constrained optimization. GAs are general purpose optimization algorithms which apply the rules of natural genetics to explore a given search space. When GAs are applied to nonlinear constrained problems, constraint handling becomes an important issue. The proposed search algorithm is realized by GAs which utilize a penalty function in the objective function to account for violation. This extension is based on systematic multi-stage assignments of weights in the penalty method as opposed to single-stage assignments in sequential unconstrained minimization. The experimental results are satisfactory and agree well with those of the gradient type methods.

Vibration Control of Flexible Structures with PZT Sensors and Actuators

In this paper, a two-degree-of-freedom model has been constructed for a structural dynamic system consisting of a linear elastic plate bonded with piezoelectric sensors and actuators. A multivariable feedback controller is designed. Four control procedures based on the minimization of performance output error and the quadratic performance index have been developed that use rate-feedback control, hybrid fuzzy-PID con trol, genetic algorithms-designed PID control, and linear quadratic Gaussian/loop transfer recovery control methods. Here, the genetic algorithm fitness function is approximately proportional to the inverse of the out put error. To test these control techniques in an efficient and systematic way, we built a digital control system that consists of MATLAB/SIMULINK modeling software and a dSPACE DS 2100 controller in a personal computer. The real-time experiment and off-line simulation results confirm that these four kinds of control methods are reliable and efficient in suppression of the steady-state resonance vibrations.

Full design of fuzzy controllers using genetic algorithms

This paper examines the applicability of genetic algorithms in the complete design of fuzzy logic controllers. While GA has been used before in the development of rule sets or high performance membership functions, the interdependence between these two components dictates that they should be designed together simultaneously. GA is fully capable of creating complete fuzzy controllers given the equations of motion of the system, eliminating the need for human input in the design loop. We show the application of this new method to the development of a cart controller.

Multiobjective evolutionary path planning via fuzzy tournament selection

The paper introduces a new selection algorithm that can be used for evolutionary path planning systems. This new selection algorithm combines fuzzy inference along with tournament selection to select candidate paths (CPs) to be parents based on: (1) the Euclidean distance from origin to destination, (2) the sum of the changes in the slope of a path, and (3) the average change in the slope of a path. The authors provide a detailed description of the fuzzy inference system used in the new fuzzy tournament selection algorithm (FTSA) as well as some examples of its usefulness. They use 12 instances of the FTSA to rank a population of CPs using the above criteria. Based on its path ranking capability, they show how the FTSA can obviate the need for the development of an explicit multiobjective evaluation function. Finally, they use the FTSA to enhance the performance of an existing evolutionary path planning system called GEPOA.

Feedback and feedforward control law for a ship crane with Maryland rigging system

A state-space model of the shipboard crane based on implicit description of the crane with no simplifying assumptions is developed. By choosing appropriate states, full control authorities for changing the length of the rope is achieved. Three control actions are considered: changing the luffing angle and changing the length of the rope from two points on the boom. A chaotic rolling moment, with a dominant frequency of the same order as the resonance frequency of the shipboard crane, is applied to the ship as an external disturbance. The effect of this disturbance is studied. Feedforward and feedback controllers are then designed and tested for this shipboard crane to suppress the load pendulation caused by ship rolling. The friction in the pulley is assumed negligible. The simulation results based on these controllers show more than 98% decrease in the pendulation magnitude due to control.

Artificial potential field based robot navigation, dynamic constrained optimization and simple genetic hill-climbing

The authors show a relationship between artificial potential field (APF) based motion planning/navigation and constrained optimization. They then present a simple genetic hill-climbing algorithm (SGHC) which is used to navigate a point robot through an environment using the APF approach. The motivation for the research is a robot that they are currently developing, named AGIE-3 (Autonomous Guided Intelligent Equipment 3), which senses and navigates through the use of a stereo vision head. They compare the SGHC with steepest descent hill-climbing (SDHC), using two environments. The first environment is composed of stationary obstacles while the second environment is composed of non-stationary obstacles. In SDHC, candidate moves are evaluated within a 360 degree radius and the best candidate is selected by the robot. One would think that the SGHC would be at a disadvantage; however, the performance of the SGHC is comparable with SDHC even though it does not search 360 degrees for candidate moves. The SGHC has an advantage in that it is capable of evolving the appropriate step size as well as the appropriate angle of movement.

Pendulation suppression of a shipboard crane using fuzzy controller

We derive the nonlinear equations of motion for a shipboard crane equipped with the Maryland Rigging. We then develop a state-space model of the crane from an implicit description without simplifying assumptions. A chaotic rolling moment with a dominant frequency of the same order as the resonance frequency of the shipboard crane is applied as an external disturbance. The effect of the disturbance is studied. A fuzzy controller is then designed and tested for this shipboard crane. In this fuzzy controller the change in the length of the rope is the control action, while the friction in the pulley is assumed negligible. The results for this controller show a big decrease in the pendulation magnitude as compared to the cases with no control.

Genetic algorithms solution for unconstrained optimal crane control

Crane control is a difficult problem for conventional control methods because of the highly nonlinear equations that must be satisfied. Usually the necessary conditions for solving an optimal control problem require finding the initial co-state vector. In this paper real-coded genetic algorithms are used to find the desired initial value of the costates of the system with no constraints. In our genetic representation, each chromosome represents a set of co-states and each gene (co-state) has an associated cost based on its ability to move the system to desired state after a given amount of time. The objective is to evolve a minimum cost co-state. Our results for this unconstrained crane problem are quite encouraging.

The role of "hierarchy" in the design of fuzzy logic controllers

This paper investigates the role of hierarchy in the systematic approach to the design of fuzzy logic controllers (FLCs). The key concept here is that the implementation of fuzzy engines at higher levels of the control hierarchy (where more reasoning is involved) yields more versatile fuzzy controllers with generally fewer control rules. At the same time, the structured nature of a hierarchical approach considerably simplifies the design procedure.