The design of an electric circuit for a certain application may be viewed as a search for solution(s) within a feasible design space formed by constraints pertaining to that application. If the feasible design space is large, it is difficult for a human designer to cover all alternatives within the feasible space to determine the best solution for a given set of criteria. If the feasible space is very small, the human designer may have difficulty in finding the few solutions available. The design process generally involves the generation of feasible alternatives, the analysis of these alternatives, and the selection of the most appropriate one(s) based on a set of criteria. Usually, due to limitation of time and resources, decisions need to be made before alternatives can be studied or even explored. A circuit designer’s knowledge and experience influences the choice of the alternatives to be pursued.
With the advent of computer technology, computer-aided design (CAD) tools have been developed to facilitate the design process and to enhance the productivity of circuit designers. Most of these CAD tools perform tasks such as optimization or analysis on a circuit given as program input. These tools are useful for the study of each individual alternative, the structure of which needs to be supplied by the designer. Tools that actually synthesize circuits are comparatively much fewer in number, and tools that generate circuit topologies and enhance the search over the feasible design space are very rare.
The objective of this paper is to demonstrate the feasibility of using computers to perform the generation of alternative circuits (that have different topologies), the evaluation of these alternatives, and the selection of the final design. A circuit design problem is recast into a search process to take advantage of the computer’s ability to search tirelessly and exhaustively over the feasible design space. Compared to other CAD tools that emphasize design analysis aspects, this approach raises the extent of design automation to a higher level by exploring and evaluating alternatives automatically.
This paper does not suggest that computers can replace humans to perform circuit designs. The objective of this is to show that it is feasible to use a computer to explore the feasible design space for a particular application so that the alternatives can be evaluated and the best solution can be selected. The usefulness of this approach increases if the circuit to be designed has a large number of alternatives. Computer programs reflective of this approach can be used both to enhance the productivity of human designers and to improve the quality of design.
With the advent of computer technology, computer-aided design (CAD) tools have been developed to facilitate the design process and to enhance the productivity of circuit designers. Most of these CAD tools perform tasks such as optimization or analysis on a circuit given as program input. These tools are useful for the study of each individual alternative, the structure of which needs to be supplied by the designer. Tools that actually synthesize circuits are comparatively much fewer in number, and tools that generate circuit topologies and enhance the search over the feasible design space are very rare.
The objective of this paper is to demonstrate the feasibility of using computers to perform the generation of alternative circuits (that have different topologies), the evaluation of these alternatives, and the selection of the final design. A circuit design problem is recast into a search process to take advantage of the computer’s ability to search tirelessly and exhaustively over the feasible design space. Compared to other CAD tools that emphasize design analysis aspects, this approach raises the extent of design automation to a higher level by exploring and evaluating alternatives automatically.
This paper does not suggest that computers can replace humans to perform circuit designs. The objective of this is to show that it is feasible to use a computer to explore the feasible design space for a particular application so that the alternatives can be evaluated and the best solution can be selected. The usefulness of this approach increases if the circuit to be designed has a large number of alternatives. Computer programs reflective of this approach can be used both to enhance the productivity of human designers and to improve the quality of design.