Numerical Methods

Report on the Spanish Telecoms Markets



Numerical Methods

Here you will find MatLab codes for "function approximation", "numerical optimization" and "systems of nonlinear equations". These codes are part of a MatLab library that computationally solves for the equilibrium of dynamic models by using pseudo-spectral methods. Unfortunately, for the moment I can only make available online the following codes:

Function Approximation

eval_pol.m: Compute Chebyschev polynomials and Chebyschev nodes. [ .m | .txt ]

Line Search Algorithm

line_search.m: A line search descent method under improved Wolfe-Powell conditions: it imposes a more stringent two-sided test on the slope, which allows the user to restrict the set of acceptable points to a small neighbourhood of a local minimizer. Based on Fletcher (2006). [ .m | .txt | example ]

I would like to improve this line search algorithm further (if possible) for the benefit of the community, suggestions and comments are welcome.