CCICAP can perform circuit optimization in the frequency domain. A circuit optimization attempts to reduce the error between a desired circuit response and the actual circuit response by adjusting specified circuit element values.
The optimization used in CCICAP is based on a genetic algorithm whereby a 'generation' of sets of circuit element values (each member of the set is called an 'individual' and represents a circuit with specific element values) is derived from a previous 'generation' of individual sets of circuit element values. The 'genetic makeup' of each individual is simply the set of component values corresponding to that individual. Associated with each individual is an error defined as the difference between that individual's circuit response and the desired circuit response.
As each new generation is created, the individuals in the current generation are 'bred' in order to produce new individuals. In CCICAP, a user specified subset of the current generation consisting of the individuals with the lowest associated errors are used to create the next full generation. The individual with the lowest error is always carried unchanged into the next generation. New members of the next generation are obtained by randomly mixing the 'genetic makeup' (circuit element values) from randomly selected individuals in the current generation. The 'breeders' are selected so that individuals with lower associated errors are selected more often than individuals with higher errors.
Each generation consists of a user specified population size. As each generation is created, there is a user specified probability of mutations. A mutation is a random change in the value of a randomly selected circuit element in a randomly selected individual.