Solving two-side numerical optimization Problems using new approach of Gradient Descent Algorithm

Abstract

In this study, we present a new gradient descent algorithm method to find first-order iterative optimization with one variable as well as a larger number of variables. The main objective of the optimization is to improve the value of the objective function taking into account a variety of constraints. The new approach algorithm depends on the process of determining the local minimum and maximum values of the function. This approach is an effective method using this strategy in applied mathematics to reach the optimal solution. The new approach algorithm is used for constraint solving and unconstrained numerical optimization.