Design of CSD Coefficient FIR filters Using Diversified ACO

A diversification strategy of ACO (Ant Colony Optimization) for designing CSD (Canonic Signed Digit) coefficient FIR filters is proposed. Although the pheromone marking in the ACO is generally carried out to just the paths that the individuals having the better objective function value have moved. As a result, an intensification ability is made strong at the earlier iterations and thus the diversification ability is made weak inversely. The CSD coefficient FIR filter design problem has a lot of local minima, and therefore such a difficulty has to be resolved for finding the better solutions even in the later iterations. In this paper, a method for avoiding this difficulty is applied to the ACO procedure. The effectiveness of the method is shown from several design results.