Polynomial approximation technique for dynamic optimization problems (Q1091280): Difference between revisions

The authors consider the general dynamic optimization problem: \[ Min\int^{t_ 1}_{t_ 0}G(x;u,t)dt\quad subject\quad to\quad \frac{dx}{dt}=h(x;u,t)\quad and\quad x(t_ 0)=x_ 0 \] and propose a polynomial approximation technique (PAT) for its solution. The authors assert that all the three approaches for solving the above problem, namely, the classical variational method, dynamic programming technique and the discretization of the problem are difficult to implement on practical problems for various reasons. The basis of PAT discussed in this paper is Ritz's method which is used to transform the given problem to an approximation in n dimensions. The well-known nonlinear programming algorithms are used to determine the optimal or near-optimal solution to the given problem. The advantage of PAT is illustrated with an example of production and inventory smoothing problem.