A new genetic single-layer routing algorithm for analog transistor arrays

In this paper a new detailed single-layer routing method for analog layout using the transistor array design style is presented. The method aims to maximize the number of routed nets and to satisfy constraints given in analog layout. A new genetic algorithm has been developed to solve this combinatorial optimization problem. For achieving good results, a new representation scheme for the individuals based on a double chromosome set is chosen. The first chromosome determines the order in which the connections are done and the second chromosome defines their shape. New efficient genetic operators to treat the single-layer routing problem have been investigated to allow a 100%-routing of all needed connections. In several tests a simulated annealing variant was not able to route some resting nets, while the implemented genetic algorithm showed much better performance. The new genetic single-layer routing approach has been tested with industrial design and the result is given