Determinati9n of approximate solution to ordinary or partial differential equations and integral equations, after discretization, leads to solving system of linear equations. Some of these equations are ill-conditioned and we have to use special methods to obtain a satisfactory approximation.
In this paper instead of solving the system
AX=B
having known bounds on the components of the vector X, it is shows that approximate solution could be found by solving the following optimization problem:
Min. ║ AX-B ║
S. t. : │xi│ ≤ δi , i=1,…,n
It is shown that this problem in well-conditioned and gives a satisfactory approximate solution to the original problem. Numerical examples are also provided to demonstrate the method.