[Paper Review] Complexity analysis of the Controlled Loosening-up (CLuP) algorithm

Mihailo Stojnic|arXiv (Cornell University)|Sep 3, 2019

Sparse and Compressive Sensing Techniques11 references4 citations

TL;DR

This paper provides a rigorous Random Duality Theory-based complexity analysis of the Controlled Loosening-up (CLuP) algorithm for MIMO maximum-likelihood detection. It demonstrates that CLuP achieves near-optimal performance in just 10 or fewer iterations across diverse problem dimensions, establishing its exceptional computational efficiency and structural simplicity in polynomial time.

ABSTRACT

In our companion paper \cite{Stojnicclupint19} we introduced a powerful mechanism that we referred to as the Controlled Loosening-up (CLuP) for handling MIMO ML-detection problems. It turned out that the algorithm has many remarkable features and one of them, the \emph{computational complexity}, we discuss in more details in this paper. As was explained in \cite{Stojnicclupint19}, the CLuP is an iterative procedure where each iteration amounts to solving a simple quadratic program. This clearly implies that the key contributing factor to its overall computational complexity is the number of iterations needed to achieve a required precision. As was also hinted in \cite{Stojnicclupint19}, that number seems to be fairly low and in some of the most interesting scenarios often not even larger than $10$. Here we provide a Random Duality Theory based careful analysis that indeed indicates that a very small number of iterations is sufficient to achieve an excellent performance. A solid set of results obtained through numerical experiments is presented as well and shown to be in a nice agreement with what the theoretical analysis predicts. Also, as was the case in \cite{Stojnicclupint19}, we again focus only on the core CLuP algorithm but do mention on several occasions that the concepts that we introduce here are as remarkably general as those that we introduced in \cite{Stojnicclupint19} and can be utilized in the analysis of a large number of classes of algorithms applicable in the most diverse of scientific fields. Many results in these directions we will present in several of our companion papers.

Motivation & Objective

• To analyze the computational complexity of the Controlled Loosening-up (CLuP) algorithm for MIMO ML detection in the high-dimensional, linear regime.
• To understand why CLuP converges in remarkably few iterations despite its iterative structure.
• To provide a theoretically grounded, iteration-by-iteration characterization of CLuP’s performance parameters using Random Duality Theory (RDT).
• To validate theoretical predictions with extensive numerical simulations, showing strong agreement between theory and experiment.
• To establish the generality of the RDT-based analysis framework for broader application beyond CLuP and MIMO detection.

Proposed method

• The analysis proceeds iteratively, characterizing key algorithmic parameters (e.g., feasibility, duality gap, solution quality) at each step using Random Duality Theory (RDT).
• The method starts with a detailed analysis of the first iteration, deriving the behavior of critical performance metrics under random matrix assumptions.
• A recursive transfer mechanism is developed to generalize the first-iteration analysis to the second, and subsequently to all subsequent iterations (k to k+1).
• The algorithm is formulated as an iterative procedure where each step solves a simple quadratic program with a norm constraint and box bounds.
• The radius parameter r controls the looseness of the constraint set and is central to convergence; its role is analyzed through duality and scaling arguments.
• Numerical experiments are conducted across various problem dimensions (n, m) with α = m/n fixed, validating theoretical predictions on iteration count and convergence speed.

Experimental results

Research questions

• RQ1 What is the theoretical basis for the observed convergence of CLuP in fewer than 10 iterations, regardless of problem size?
• RQ2 How can the performance of each individual iteration in CLuP be characterized independently and systematically using Random Duality Theory?
• RQ3 To what extent can the first-iteration analysis be extended to all later iterations via a recursive transfer mechanism?
• RQ4 How well do theoretical predictions on iteration count and convergence behavior align with empirical results in numerical simulations?
• RQ5 What is the role of the radius parameter r in controlling the trade-off between convergence speed and solution quality in CLuP?

Key findings

• The number of iterations required for CLuP to achieve high-precision solutions is consistently small—often not exceeding 10, even for large problem dimensions.
• Theoretical analysis using Random Duality Theory (RDT) successfully characterizes the evolution of key performance metrics (e.g., duality gap, feasibility) across all iterations.
• A recursive transfer mechanism from iteration k to k+1 is established, enabling a general framework to analyze the algorithm beyond the first few steps.
• Numerical experiments show excellent agreement between theoretical predictions and observed convergence behavior across multiple problem regimes.
• The CLuP algorithm’s structural simplicity and low iteration count suggest it is a strong candidate for the MIMO ML grand-challenge: exact solution in polynomial time with minimal complexity.
• The analysis framework is not limited to CLuP or MIMO detection but is generalizable to a wide class of algorithms in optimization, statistics, and machine learning.