China
Africa
Explore chapters and articles related to this topic
Integration of Neural Network Techniques with Approximate Reasoning in Knowledge-Based Systems
Published in Abraham Kandel, Gideon Langholz, Lotfi A. Zadeh, Hybrid Architectures for Intelligent Systems, 2020
Rule-based expert systems have found many applications in the last two decades [1]. These systems opened new frontiers in computing in that they manipulated symbolic rather than numerical data. This approach offered a number of advantages, including the ability to develop a general purpose inference engine, or reasoning structure, which was totally separated from the knowledge base for the application [2]. The reasoning structure could then be used without change in new applications [3]. The development of the knowledge base was accomplished through consultation with domain experts. This procedure was and is a time-consuming and labor-intensive operation, especially for complex knowledge domains, although some attempts have been made to automate the process [4]. The knowledge-based approach also offered the important feature of explanation capabilities which did not previously exist. This shortcoming was in fact one of the major objections to earlier systems which utilized pattern recognition techniques. Another advantage to the knowledge-based approach to automated reasoning was the inclusion of, in many of these systems, some ability to deal with uncertain or missing information [5]. This aspect of knowledge-based systems continues to be an active area of research [6–10], with only partial solutions practically implemented at this point in time [11–15]. The major drawback to the knowledge-based approach remains the difficulty of developing the knowledge base [16,17].
Sensor and data fusion in traffic management
Published in Lawrence A. Klein, ITS Sensors and Architectures for Traffic Management and Connected Vehicles, 2017
Knowledge-based systems incorporate rules and other knowledge from known experts to automate the object identification process. They retain the expert knowledge for use at a time when the human inference source is no longer available. Computer-based expert systems frequently consist of four components: (1) a knowledge base that contains facts, algorithms, and a representation of heuristic rules; (2) a global database that contains dynamic input data or imagery; (3) a control structure or inference engine; and (4) a human–machine interface. The inference engine processes the data by searching the knowledge base and applying the facts, algorithms, and rules to the input data. The output of the process is a set of suggested actions that is presented to the end user [49].
System Design
Published in Robert F. Hodson, Abraham Kandel, Real-Time Expert Systems Computer Architecture, 1991
Robert F. Hodson, Abraham Kandel
The inference engine in a rule-based system is typically forward chaining (data driven) or backward chaining (goal driven). Real-time systems can have both goal driven and data driven characteristics. Hierarchical control and domain decomposition techniques lend themselves to a goal directed approach. The system requirement of responsiveness (to external stimuli) reflects a data driven approach. In this system, a compromise was made by using a modified backward chaining inference mechanism. Backward chaining provides a more focused and methodical approach to the task [Short77]. Backward chaining performs a depth first search of the solution space starting from the system goal and decomposes the problem into subgoals. This natural decomposition of the problem space uncovers inherent parallelism in the problem that can be exploited by the system. An extension to the backward chaining approach to improve system responsiveness is to allow the control mechanism to query the real-time environment directly for information. The system can also wait on information or the occurrence of an event to trigger inferencing.
Development of an expert system for demand management process
Published in International Journal of Computer Integrated Manufacturing, 2018
The knowledge base as seen Figure 1 is the part of an ES that contains domain knowledge which may be expressed as any combination of ‘IF–THEN’ rules, factual statements, frames, objects, procedures and cases. The inference mechanism manipulates the stored knowledge to produce solutions to problems (Pham and Pham 1999). The inference engine uses the rules to automatically determine what information is needed, the implications of various facts and arrives at a logically reasoned conclusion. It interprets knowledge in the knowledge base and makes reasoning among knowledge rules. Explanation is an attempt by an ES to clarify its reasoning, recommendations or other actions, e.g. asking a question.
Backward chaining inference as a database stored procedure – the experiments on real-world knowledge bases
Published in Journal of Information and Telecommunication, 2018
Tomasz Xie¸ski, Roman Simiński
The inference algorithm selects some applicable rules to infer new facts and/or confirm established goals. When rules are examined by the inference engine, new facts are added to the fact base if its current content satisfies the conditions in the rules. The strategy of backward chaining is started from a goal and ends with a set of facts that leads to the given goal, and therefore, it is also known as a goal-driven strategy of the inference engine.
Equivalent Salt Deposit Density Prediction of Silicone Rubber Insulators Under Simulated Pollution Conditions
Published in Electric Power Components and Systems, 2018
Abdelrahman K. Abouzeid, Ayman El-Hag, Khaled Assaleh
The inference engine is responsible for processing the inference operations on the rules. The fuzzification interface will find the degree of match between the input and their linguistic value, represented by the membership function. Finally at the defuzzification phase, the fuzzy result will be transformed back into its crisp output value.