Information Sciences - Applications最新文献

Reports on successful applications of fuzzy logic controllers (FLCs) are no longer rare. Regardless of the application domain, the main idea is to convert a linguistic control scenario into an automatic control strategy. The expert's knowledge is the backbone of this linguistic control strategy. FLCs have their most successful implementations where the process under control is too complex for analysis by conventional quantitative methods and, therefore, conventional controllers face serious shortcomings. This paper proposes a “hybrid” implementation of FLCs and conventional PID controllers that can be helpful in some applications. The proposed method is applied to a 2 degree of freedom robot arm with promising results.

This paper looks at the possibility of using fuzzy terms in a software requirements specification. Potential benefits are reviewed. Methods for testing whether a fuzzy specification is met are investigated. Ways of modeling criticality and rigidity of specifications using fuzzy approaches are also identified. An example based on query response times is used throughout the paper. The results of a pilot study to see whether computer response time requirements can be fuzzified are described.

This paper addresses the simulation of four networks whose underlying topology is a Cayley graph. The SIMD model of parallelism is used in this study. The networks considered are: star graph, bubble-soart graph, and pancake graph. The simulation is performed by mimicking the interconnection functions of the guest (network to be simulated) using a set of interconnection functions offered by the host. The interconnection function of the network to be simulated that requires the most time to simulate determines the simulation time of that network. The simulation times for all cases are derived and shown to be optimal. Results indicate the superiority of pancake networks over star networks in simulating the networks under study.

Applications of complex-valued neural networks to optical signal processing using phase-sensitive detection schemes are proposed and discussed. In optical information processing systems, the advantages of complex-valued neural networks are realized directly utilizing the physical phenomena of lightwave and phase-sensitive optical detection. In the proposed system, the summation and nonlinear transform of the complex signals are processed quickly in parallel by using an optical circuit. The phase-sensitive detection scheme is used for extracting the complex-valued signals. The complex-valued neural networks implemented as optical neural systems provide a starting point for coherent optical neural networks.

The slow and uncertain convergence of multilayer feedforward neural networks using the backpropagation training algorithm is caused mainly by the iterative nature of the dynamic process of finding the weight matrices with static control parameters. This study investigates the use of fuzzy logic in controlling the learning processes of such neural networks. Each learning neuron in the neural networks suggested here has its own learning rate dynamically adjusted by a fuzzy logic controller during the course of training according to the output error of the neuron and a set of heuristic rules. Comparative tests showed that such fuzzy backpropagation algorithms stabilized the training processes of these neural networks and, therefore, produced 2 to 3 times more converged tests than the conventional backpropagation algorithms. The sensitivities of the training processes to the variations of fuzzy sets and membership functions are examined and discussed.

This paper presents a mathematical model for optimally siting and routing hazardous waste operations conditioned on public perception toward acceptable costs and risks. The routing and siting of hazardous waste operations is governed as much by the public's perception of acceptable cost and risk as by any other factor. These perceptions are integrated into the model through the use of fuzzy sets that represent the public's degree of acceptance toward unique policy options. Perceived risk is used to help determine a policy's acceptability. Model solutions define a trade-off relationship between cost, risk, and the perception that the policy is acceptable. Initially, linear fuzzy membership functions for acceptable cost and risk are used to demonstrate the efficacy of this approach. Following the linear formulation, more realistic and computationally complex nonlinear membership functions are incorporated into the model. To illustrate the models, a case study that considers the current effort by North Carolina to site a hazardous waste incinerator is presented.

Shape recognition is an important research area in pattern recognition. It also has wide practical applications in many fields. An attribute grammar approach to shape recognition combines both advantages of syntactic and statistical methods and makes shape recognition more accurate and efficient. However, the time complexity of a sequential shape recognition algorithm using attribute grammar is O(n³) where n is the length of an input string. When the problem size is very large, it needs much more computing time; therefore, a high-speed parallel shape recognition algorithm is necessary to meet the demands of some real-time applications. This paper presents a parallel shape recognition algorithm, and also discusses the algorithm partition problem as well as its implementation on a fixed-size VLSI architecture. The proposed algorithm has time complexity O(n³/k²) if using k × k processing elements. When k = n, its time complexity is O(n). The experiment has been conducted to verify the performance of the proposed algorithm. The correctnes of the algorithm partition and the behavior of proposed VLSI architecture have also been proved through the experiment. The results indicate that the proposed algorithm and the VLSI architecture could be very useful to imaging processing, pattern recognition, and related areas, especially for real-time applications.