Using Multivalued Cartesian Genetic Programming (M-CGP) for Automatic Design of Digital Sequential Circuits

Abstract

The paper addresses the problem of the automatic design of sequential systems. For a complete description of the operation of the sequential system, a table of states or another representation of transition graphs describing possible changes in system states is necessary. This paper adopts a completely different approach, in which the description of the sequential system results from the study of the responses to signals given from outside and from an unknown system, which is treated as a black box. This approach may be useful when we want to recreate the internal structure of a given, unknown system or when we want to obtain a system based only on the information about the system’s reactions to given external signals, without going into the principles of its operation. The paper presents problems that arise when creating the data strings that describe the reactions of the designed system and ways for solving these problems, and it presents Multivalued Cartesian Genetic Programming (M-CGP)—a new approach used to design sequential circuits. Further research has developed a system based on this model. The paper presents examples of obtained sequential systems generated using the newly created system.

Keywords

• genetic programming
• cartesian genetic programming
• CGP
• sequential circuits
• sequential logic circuit
• M-CGP