cuikgradient.c File Reference Detailed DescriptionFind a solution of a cuiksystem set of equations using a gradient descent method. It is based on an iterative process starting from a random point sampled in the initial box.
The gradient descent method minimizes an error function defined as
Observe that the gradient of this function is
Which can be expressed as
Where J_f^t is the transposed of the Jacobian of the initial system and (f_1....f_n) is the evaluation of this system.
Definition in file cuikgradient.c.
Function Documentation
Main body of the cuikgradient application. Use:
where
In this case we sample points at random in the initial box and try to converge to the manifold. If the gradient process succeeds, we print the solution point, a tiny box around it, and the error in the solution point. The program can also be used as:
where
In this case we search for a solution file and we try to converge to a point for each one of the boxes in the solution file. If the solution file does not exists an error is triggered. In any cases .cuik extension is optional.
Definition at line 95 of file cuikgradient.c. References CONVERGED_IN_BOX, CONVERGED_IN_GLOBAL, CONVERGED_OUTSIDE_GLOBAL, CreateFileName(), CT_REPRESENTATION, CUIK_EXT, CuikGradient(), CuikGradientInBox(), DeleteBox(), DeleteCuikSystem(), DeleteFileName(), DeleteParameters(), DIVERGED, Error(), ErrorInCSEquations(), ErrorInSolution(), GetCSSystemVars(), GetFileFullName(), GetParameter(), InitCuikSystemFromFile(), InitParametersFromFile(), JOINTS_EXT, LINKS_EXT, NEW, PARAM_EXT, PrintBoxSubset(), randomSet(), ReadBox(), REP_JOINTS, and SOL_EXT. |
Follow us!