A Unified MIMO Optimization Framework Relying on the KKT Conditions

A popular technique of designing multiple-input multiple-output (MIMO) communication systems relies on optimizing the positive semidefinite covariance matrix at source. In this paper, a unified MIMO optimization framework based Karush-Kuhn-Tucker (KKT) conditions is proposed. framework, with aid theory, Theorem 1 presents generic optimal transmit for diverse objective functions subject to various power constraints and different levels channel state information (CSI). Specifically, fundamentally reveals that family systems, matrices associated under can be derived in water-filling-like form. When applying case multiple general constraints, we firstly equivalently transform into single counterpart by introducing weighting factors Pareto theory. The found proposed modified subgradient method. On other hand, imperfect system statistical CSI errors, address non-convexity robust problem following idea alternating optimization. Finally, our numerical results verify solution structure global optimality method, as well demonstrate performance advantages algorithm.