Interior point column generation algorithms for adaptive filtering

Kaywan H. Afkhamie, Zhi Quan Luo, Kon Max Wong

Research output: Contribution to journalArticlepeer-review


This paper presents a novel application of interior point column generation (IPCG) algorithms to adaptive Altering. IPCG algorithms have been developed in optimization theory to solve convex optimization or convex feasibility problems. One of their key features is that inequality constraints (that define the convex feasible region) are considered one at a time. This feature can be exploited in applications that require a solution to be updated adaptively. We give some background on IPCG algorithms and how they are used to solve convex feasibility problems. Then, we apply IPCG to two classical filtering problems: adaptive channel equalization and adaptive system identification. Our simulation results show that under adverse conditions (e.g., low SNR's, correlated input signals, and time-varying systems), the IPCG exhibits convergence behavior superior to the recursive least squares (RLS) method, whereas under less adverse conditions, the IPCG algorithm consistently matches the RLS performance.

Original languageEnglish (US)
Pages (from-to)2627
Number of pages1
JournalIEEE Transactions on Signal Processing
Issue number10
StatePublished - 1997


Dive into the research topics of 'Interior point column generation algorithms for adaptive filtering'. Together they form a unique fingerprint.

Cite this