Optimizing Gradient Methods for IoT Applications

Solving linear programming (LP) and nonlinear (NLP) problems is momentous because of their wide applications in real-life problems. There no unified way to find the global optimum for NLPs. But on other hand, simplex algorithm, as dominating methodology LPs several decades moves only boundary (vertices) ignores vast majority feasible region process searching. In this article, we study two gradient-based methodologies that explore whole region, which guarantee faster convergence rates both LP NLP optimization including IoT such Software-Defined Internet Vehicles (SDIoV) vehicular ad hoc networks (VANETs). The gradient-simplex algorithm (GSA) LPs, inside gradient direction at first reduce search space then explores reduced an optimal solution. evolutionary-gradient (EGA), NLPs uses evolutionary population estimate gradients by evolving better solutions steps. Based extensive simulations, obtained numerical results show approaches provide efficient outperform state-of-the-art methods with large spaces. Comparative applying GSA SDIoV VANETs different sizes are included.