State simplification method of sequential logic circuit based on granular computing

Abstract

The invention discloses a granular computing-based sequential logic circuit state simplification method. According to the method, a next state matrix for describing state transition conditions is defined, and the next state matrix is continuously updated through marking compatible classes, so that a maximum compatible class set is solved and then the generation of a large-scale sparse matrix is avoided; in the process of solving maximum compatible classes, all the initial states in a state transition table are directly divided, and through iteration, a final result can be obtained, so that theprocess of solving compatible pairs and incompatible pairs of the initial states in other algorithms is avoided and then the time expenditure is decreased; by taking kernel compatible classes as heuristic information to construct an initial state tree, all the possible minimum coverages can be solved rapidly; and through constructing a minimum state tree, closing performance of all the minimum coverages can be verified and a simplification result with minimum states can be obtained, so that the optimality of the algorithm result is ensured.

Description

technical field [0001] The invention relates to the field of circuit simplification, in particular to a granular computing-based sequential logic circuit state simplification method. Background technique [0002] The good design of sequential logic circuit mainly depends on the accuracy and simplicity of the state transition table (graph) established according to the problem. Accuracy refers to expressing the state changes and transitions of input, output, and circuits in a graph in strict accordance with specific problems; simplicity refers to eliminating redundant states in the table as much as possible, because the fewer the number of states, the more memory circuits. Simpler, higher system reliability. [0003] State simplification is the process of eliminating redundant states in the original state transition diagram and obtaining the minimum state transition table. According to whether the output and the next state have uncertain values, sequential circuits can be di...

AI Technical Summary

Problems solved by technology

Zhang Xianda uses the simplified hidden graph of the maximum compatible class, uses the recursive selection of trap points and circuit decomposition to obtain the minimum closed cover. This method is a graphical method for obtaining the minimum closed cover. In the process, the traditional implicit table method is still used in essence. It is necessary to find all compatible state pairs, incompatible state pairs and implicitly compatible state pairs from the state transition table, and find the minimum closed coverage in the decomposition loop In the process, it is often impossible to succeed at one time, and many experiments are required, which leads to a great complexity of the algorithm; at the same time, someone proposed a fast algorithm for an incompletely determined sequence machine. A small number of "candidate sets", by continuously modifying this "candidate set", the members of the minimum closed coverage are gradually introduced into the "candidate set", so as to obtain the desired result

However, this algorithm cannot guarantee that the obtained result must be a minimum closed cover, but is often a quasi-minimal closed cover close to the minimum closed cover, resulting in insufficient accuracy of the results. In addition, the algorithm is not formal enough and is not easy to program; Higuchi, H. and Matsunaga, Y. proposed an improved iterative algorithm in which a binary decision graph was used to avoid the combinatorial explosion problem; Goren, S. et al proposed a heuristic state reduction algorithm based on branch and bound Search technology and identification technology improve the efficiency of reduction. By first finding the upper and lower bounds of the state transition table, and then using the search algorithm, gradually construct the results that meet the requirements, but in the process of demarcation and search, it is necessary to build incrementally Input / output sequence, and continuously introduce additional states for identification, which increases the size of the original state table and increases storage overhead

Images