Non-convex optimization techniques for real-time signal representation using hybrid algorithms

Non-Convex Optimization Techniques; Real-time signal processing is important in many engineering fields such as communication systems, biomedical engineering, and audio/video signals. However, the major challenge lies in finding an optimal signal representation in real-time scenarios owing to the highly non-linear nature of the optimization problem. In this paper, we show a hybrid approach by taking advantage of convex and non-convex optimization to tackle the issues of real-time signal representation. A hybrid solution uses a multi-objective, non-convex optimization approach combined with a convex-like optimization technique for an exact and efficient representation of the underlying signal. Non-convex optimization paves the way for more expressivsignal modeling, while convex counterpart leads to efficient computation. The proposed method is tested on real data and compared with some classic methods, proving better accuracy and computation speed. The suggested hybrid algorithm can fulfill the needs of other real-time signal processing applications, contributing to this domain.