Optimal Rule Base Research Articles

Identification of protein pattern in kidney cancer using ProteinChip® arrays and bioinformatics

Tumor biology of renal cell carcinoma (RCC) is not very well understood, although many studies on molecular and cellular biology have been performed. It is accepted now that cancer research has to be performed also with proteomic tools, because proteins are the real actors in the genesis and progression of cancer. Therefore, we used a ProteinChip System(R) (SELDI) which is able to detect minute amounts of protein and moreover to analyze a complex protein pattern. We analyzed 37 cases of clear cell RCC as a training set including corresponding normal tissue. From all samples protein lysates were made and spotted directly on different chip surfaces (SAX2, WCX). After a washing procedure the arrays were analyzed in the ProteinChip Reader. All profiles were subjected to a bioinformatical analysis including normalization, clustering, rule extraction and rating. Defined rules (markers) were evaluated using a test set of 24 samples (13 tumor tissues and 11 normal kidney tissues). The generated rule base for the SAX2 surface showed a sensitivity of 100% and a specificity of 97.3%. For the WCX arrays the optimal rule base showed worse results. A combined rule base for SAX2 and WCX did not result in a higher sensitivity or specificity. Using the optimal rule base for the SAX2 chip in the test set, sensitivity and specificity reached 76.9% and 100%, respectively. The ProteinChip System represents a key technology for the rapid detection of cancer specific proteomic patterns. It is possible to identify clear cell renal cancer with high sensitivity and specificity from minimal amounts of cells.

Intelligent control planning strategies with neural network/fuzzy coordinator and sensor fusion for robotic part macro/micro-assembly tasks in a partially unknown environment

This paper addresses the problems of the robotic part assembly in a partially unknown environment. The process can be broken down into two phases. First, a macro-assembly, locating various shaped assembly holes or receptacles (targets) in the workspace corresponding to the shapes of the parts and then bringing the part to the corresponding target, despite existing obstacles. This is accomplished by combining a neural network control strategy coordinating with a mobile multiple optical sensor as well as fuzzy optimal controls. Depending on topological relationships among the part’s present position, the position of obstacles, and the target position in the workspace, a specific rulebase from a family of distinct fuzzy rulebases for avoiding obstacles is activated. An entropy function, which is a useful measure of the variability and the information in terms of uncertainty, is introduced to measure its overall performance of a task execution related to the part assembly task. Second, a micro-assembly, placing the part at a position that is ready to mate successfully with the target without jamming. Depending on a mating type, namely, a rightside, a leftside, or a straight approach toward a target, which is determined by fusing sensor information obtained by optical sensors, a specific rulebase is activated. Fuzzy set theory is used to address the uncertainty associated with the macro and the micro-assembly procedures. Using the control of a robotic part assembly task as an example, a systematic method, not a heuristic one, that can determine an optimal rulebase among feasible fuzzy rulebases which can execute the part assembly task successfully, based on a fuzzy entropy is introduced. The degree of uncertainty associated with the part assembly tasks is used as an optimality criterion, e.g. minimum fuzzy entropy, for a specific task execution. The results show the effectiveness of the above methodologies. The proposed technique is applicable to a wide range of robotic tasks including part mating with various shaped parts, pick and place operations, and motion planning.

An expert system based on FBFN using a GA to predict surface finish in ultra-precision turning

Ultra-precision turning is an important operation used to generate a high surface finish in precision component and its input–output relationships are highly non-linear. Surface finish of a turned surface depends on the selection of cutting parameters, such as cutting speed, feed and depth of cut. But, it is difficult to obtain the exact mathematical expression for surface finish in terms of the input variables. Realizing the fact that fuzzy logic controller (FLC) is a potential tool for dealing with impression and uncertainty, an attempt was made by the authors previously to model the grinding operation using a combined GA–fuzzy approach. In that approach, an optimal FLC was developed using a GA-based tuning. But, the main drawback of that approach lies in the fact that it does not keep track on the number of times a particular rule is getting fired during training. Thus, the resulting optimized rule base (RB) of an FLC may contain some redundant rules. To overcome this, an expert system is developed in this paper, based on the fuzzy basis function network (FBFN) to predict surface finish in ultra-precision turning. An approach for automatic design of RB and the weight factors (WFs) for different rules is developed using a GA, based on error reduction measures. Results of the proposed expert system have been compared to the experimental results available in the literature.

Diagnosis of the diseases––using a GA-fuzzy approach

The objective of our study is to design an expert system by modelling the knowledge and thinking process of a doctor. A fuzzy logic controller (FLC) is used to model the process and a genetic algorithm (GA) helps to select a number of good rules from a manually constructed large rule base of an FLC, based on the opinion of 10 doctors. The GA-based tuning is done off-line. Once the optimized rule base of the FLC is obtained, it can diagnose the disease, on-line. The scope of the present work has been extended to two diseases, namely Pneumonia and Jaundice. The symptoms of each disease are fed as inputs to the FLC and the output, i.e., grade of a disease is determined.

Function estimation using a neural-fuzzy network and an improved genetic algorithm

This paper presents the estimation of the transmission gains for an AC power line data network in an intelligent home. The estimated gains ensure the transmission reliability and efficiency. A neural-fuzzy network with rule switches is proposed to perform the estimation. An improved genetic algorithm is proposed to tune the parameters and the rules of the proposed neural-fuzzy network. By turning on or off the rule switches, an optimal rule base can be obtained. An application example will be given.

On-line fuzzy identification using genetic algorithms

The design and implementation of an on-line fuzzy identification method using genetic algorithms (GAs) is reported in this paper. In the proposed algorithm, the rule-table of a fuzzy system is first divided into several independent and much smaller fuzzy systems, which in turn are encoded into separate bit strings for genetic operations. A novel GA updating architecture is then proposed to search the optimal rule-base of these fuzzy systems at each sample interval. The performance of this identification algorithm is evaluated by simulation of a non-linear system. Moreover, an experiment on simple behavior learning of a mobile robot is also reported. The results indicate an improvement in design cycle and convergence to the optimal rule-base within a relatively short period of time.

Predictive fuzzy PID control: theory, design and simulation

Predictive fuzzy PID control theory is developed in this paper, which offers a new approach for robust control of time-delay systems. The paper describes the functional structure, design principle, and stability analysis of a new predictive fuzzy PID controller. Sufficient computer simulations are provided for illustration and verification. First, the structure of the controller is derived from both the fuzzy PID control and the generalized predictive control concepts. Then, on-line model identification, optimal cost index, fuzzification , rule base, and defuzzification of the representative predictive fuzzy PD+I controller are discussed in detail. Lyapunov asymptotic stability analysis is conducted. Then, many computer simulations are performed to compare with several closely related controllers such as the fuzzy PD+I controller and the Smith-type predictive fuzzy PD+I controller. In the simulations, second-order linear systems with/without time delays, nonlinear systems with/without time delays, uncertain linear systems with time delays, and uncertain nonlinear systems with time delays are used to confirm the advantages of the new predictive fuzzy PD+I controller. Finally, this method is applied to control some chaotic systems with success. This predictive fuzzy control method provides a new way for controlling uncertainty and complex linear and nonlinear systems, even with significant time delay.

An optimal design of neuro-FLC by Lamarckian co-adaptation of learning and evolution

This paper proposes a new design method of neuro-FLC by the Lamarckian co-adaptation scheme that incorporates the backpropagation learning into the GA evolution in an attempt to find optimal design parameters (fuzzy rule base and membership functions) of application-specific FLC. The design parameters are determined by evolution and learning in a way that the evolution performs the global search and makes inter-FLC parameter adjustments in order to obtain both the optimal rule base having high covering value and small number of useful fuzzy rules and the optimal membership functions having small approximation error and good control performance while the learning performs the local search and makes intra-FLC parameter adjustments by interacting each FLC with its environment. The proposed co-adaptive design method produces better approximation ability because it includes the backpropagation learning in every generation of GA evolution, shows better control performance because the used COG defuzzifier does compute the crisp value accurately, and requires small workspace because the optimization procedure of fuzzy rule base and membership functions is performed concurrently by an integrated fitness function on the same fuzzy partition. Simulation results show that the Lamarckian co-adapted FLC produces the most superior one among the differently generated FLCs in all aspects such as the number of fuzzy rules, the approximation error, and the average tracing distance.

Evolutionary design of fuzzy rule base for nonlinear system modeling and control

In designing fuzzy models and controllers, we encounter a major difficulty in the identification of an optimized fuzzy rule base, which is traditionally achieved by a tedious trial-and-error process. The paper presents an approach to the evolutionary design of an optimal fuzzy rule base for modeling and control. Evolutionary programming is used to simultaneously evolve the structure and the parameter of fuzzy rule base for a given task. To check the effectiveness of the suggested approach, four numerical examples are examined. The performance of the identified fuzzy rule bases is demonstrated.

TEMPERATURE INFERENCE SYSTEM BY ROUGH-NEURO-FUZZY NETWORK

Rough set theory, suggested by Pawlak in 1982, has been useful in AI, machine learning, knowledge acquisition, knowledge discovery from databases, expert systems, inductive reasoning, etc. One of the main advantages of rough sets is that they do not need any preliminary or additional information about data and are capable to reduce superfluous information. However, their significant disadvantage in real applications is that inferences are not real values but disjoint intervals of real values. In order to overcome this difficulty, we propose a new approach in which rough set theory and neuro-fuzzy fusion are combined to obtain the optimal rule base from input/output data. These results are applied to the rule construction for inferring the temperature at specified points in a refrigerator.

A genetic-fuzzy approach for mobile robot navigation among moving obstacles

In this paper, a genetic-fuzzy approach is developed for solving the motion planning problem of a mobile robot in the presence of moving obstacles. The application of combined soft computing techniques — neural network, fuzzy logic, genetic algorithms, tabu search and others — is becoming increasingly popular among various researchers due to their ability to handle imprecision and uncertainties that are often present in many real-world problems. In this study, genetic algorithms are used for tuning the scaling factors of the state variables (keeping the relative spacing of the membership distributions constant) and rule sets of a fuzzy logic controller (FLC) which a robot uses to navigate among moving obstacles. The use of an FLC makes the approach easier to be used in practice. Although there exist many studies involving classical methods and using FLCs they are either computationally extensive or they do not attempt to find optimal controllers. The proposed genetic-fuzzy approach optimizes the travel time of a robot off-line by simultaneously finding an optimal fuzzy rule base and optimal scaling factors of the state variables. A mobile robot cant then use this optimal FLC on-line to navigate in presence of moving obstacles. The results of this study on a number of problem scenarios show that the proposed genetic-fuzzy approach can produce efficient knowledge base of an FLC for controlling the motion of a robot among moving obstacles.

Automatically extracting and optimizing fuzzy rule base

A method of automatically extracting and optimizing fuzzy rule base is presented in this paper. Firstly, it applies the method of CCM (cell-to-cell mapping) to analyze the evolving trend and global behavior of a fuzzy dynamical system based on a cell state space, which is characterized by equilibrium, cost and their domain of attractions. Secondly, each rule base is evaluated to determine a performance index based on the information of the system obtained by CCM. Thirdly, GA (Genetic Algorithm) optimizes the coded rule bases according to the performance index generation by generation. The method presented in this paper can be applied to various systems (linear or nonlinear, continuous or discrete) to automatically obtain optimal rule base, for it fuses the advantages of GA and CCM. As an example, a complicated nonlinear system—an inverted pendulum is simulated to demonstrate the validity of the method.