Research On Automatic Synthesis Of Combinational Logic Circuits

The design of combinational logic circuits is one of the most important aspects in the digital circuit design. Traditional design methods need that designers have rich theory foundation and design experience. Although the designers can design satisfied circuits, the obtained circuits are not always optimal. Therefore, some effective approaches are proposed to overcome the disadvantage of traditional design methods. Among these methods, the automatic synthesis of combinational logic circuits has shown significant advantages, it can bring breakthrough to the field of digital circuit design.In this paper, an improved gene expression-based clonal selection algorithm (IGE-CSA) is proposed firstly, which is applied into the automatic design of combinational logic circuits. In the circuit encoding, we improve the encoding of gene expression programming in two aspects. On one hand, we use the concept of block set to shorten the length of the circuit encoding. On the other hand, the concept of output region is proposed so that multi-output combinational logic circuits can be represented by one gene encoding. Combined with the proposed encoding, clonal selection algorithm (CSA) is utilized as search engine so that our proposed method has more powerful global search ability. Then, another novel method --- graph-based clonal selection algorithm (GCSA) is proposed to synthesize combinational logic circuits automatically. In this method, we propose the graph encoding to represent combinational logic circuits. The graph encoding can represent any combinational logic circuits and mapping rules from graph to circuit are very simple. CSA is also applied into this method. In CSA method, eight innovative mutation operators are designed. Being different from conventional genetic algorithm, genetic programming and gene expression programming, these operators can directly manipulate the graph structure so that the variety of graph is very rich.The proposed two methods are applied into combinational logic circuit design successfully. Finally, four kinds of combinational logic circuits are synthesized to verify the effectiveness of the proposed two approaches. The experimental results show that the proposed approaches can not only automatically design satisfied circuits but also the obtained circuits might be optimal or close to optimal. Furthermore, compared with another five methods, the experimental results show that the proposed approaches have the highest efficiency and success rate.