I-flows: A novel approach to computational intelligence for analog circuit design automation through symbolic data mining and knowledge-intensive reasoning

This chapter presents an overview of the authors’ ongoing work toward devising a new approach to analog circuit synthesis. The approach computationally implements some of the facets of knowledge-intensive reasoning that humans perform when tackling new design problems. This is achieved through a synthesis flow that mimics reasoning using a domain-specific knowledge structure with two components: an associative part and a causal reasoning part. The associative part groups known circuit schematics into abstractions based on the similarities and differences of their structural features. The causal reasoning component includes the starting ideas as well as the design sequences that create the existing circuits.