A Graph-Based Approach for Timing Analysis and Refinement of OPS5 Knowledge-Based Systems
IEEE Transactions on Knowledge and Data Engineering
Optimizing Real-Time Equational Rule-Based Systems
IEEE Transactions on Software Engineering
| Hi-index | 0.00 |
Real-time decision systems (RTDS''s) are computer-controlled systems that must react to events in the external environment by performing decision-intensive computation sufficiently fast to meet specified timing and safety constraints. This dissertation investigates a class of these systems where decisions are computed by an equational rule-based program. Two fundamental problems are identified: (1) the analysis of rule-based RTDS''s in order to verify that the specified timing and safety properties are satisfied prior to their execution, and (2) the synthesis of rule-based RTDS''s that are guaranteed to meet the specified timing constraints in addition to the safety constraints. Two complementary approaches have been developed to solve the first problem: (1) model checking of the global state transition graph representing the program, and (2) static analysis of the program. These approaches are combined to form the cornerstone of the General Iterative Analysis Algorithm. The applicability of this analysis technique is further enhanced by the development of a facility with which the rule-based programmer can specify domain-specific knowledge in the language Estella in order to validate the performance of an even wider range of programs. Two approaches also have been identified to tackle the second problem: (1) transforming the given equational rule-based program by adding, deleting, and/or modifying rules, and (2) optimizing the scheduler to select the rules to fire such that the variables in the program will always converge to stable values within the response time constraint. The complexity and size of real-time decision systems often necessitates the use of computer-aided design tools. This dissertation describes a suite of analysis tools based on our theoretical framework which have been implemented to ensure that equational rule-based programs written in the language EQL can indeed meet their specified timing constraints.