Design of FIR Filters with Discrete Coefficients Using Ant Colony Optimization

In this paper, we propose a new design method for linear phase finite impulse response (FIR) filters with discrete coefficients. In a hardware implementation, filter coefficients must be represented as discrete values. The design problem of digital filters with discrete coefficients is formulated as an integer programming problem. But an enormous amount of computational time is required to solve the problem in a strict solver. Recently, ant colony optimization (ACO), a heuristic approach, has been used widely for solving combinational problems such as the traveling salesman problem. In our method, we formulate the design problem as a 0-1 integer programming problem and solve it by using ACO. Several design examples are presented to show the effectiveness of the proposed method.