Forums

Design FIR filters using DRHS-OLL algorithm

Started by jtp_1960 3 weeks ago1 replylatest reply 3 weeks ago102 views

Bumped to research paper Efficient FIR Filter Design Methodology Using Dynamic Regional Harmony Search Algorithm With Opposition And Local Learning (DRHS-OLL) written by P.Nirmala and Dr. R. Rani Hemamalini.

Research paper for DRHS-OLL algorithm written by A.K. Qin and Florence Forbes

My math/DSP skills are not good enough for coding algorithms/equations given in those papers so, have to ask if someone reading this post has already tried this method in FIR filter design (maybe open sourced matlab, c++, c, python, etc.) or can point some implementation with source codes (github, sourceforge, etc.)?

By "after reading the papers" or practice, would there be advantages/disadvantages, compared to commonly used design methods (windowing, Parks-McClellan, IRLS), in audio related applications?

There's also recent attempt to use same method in design of an IIR (LPF) filter.

[ - ]
Reply by Y(J)SAugust 26, 2021

Filter design is an optimization problem, and as such many optimization methods may be applied.

HS-based optimization is yet another attempt at producing an optimization method based on an analogy. So far, it has not produced anything out of the ordinary, and in the paper you quote it underperforms well-known methods.

If you really want to design a FIR filter, 99 times out of 100 Remez is fine. If you want to experiment with optimization heuristics use simulated annealing which is more successful in general than analogy-based methods that are basically pruning exhaustive search.