Cubic Spline calculation after optimization

1 次查看（过去 30 天）

David 2014-10-21

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/159490-cubic-spline-calculation-after-optimization

Hello dear MATLAB community,

at the moment I try to calculate a trajectory by solving a parametrised optimization problem. This works pretty good for many cases. But there are some cases where I get weird results.

The picture shows a weird calculated trajectory. The spline should go through the points but its shape is very weird. At the moment I´m calculating the trajectory with :

F1 = griddedInterpolant(f,x,'spline');
Way(1,:) = F1(x2);
F3 = griddedInterpolant(f,y,'spline');
Way(2,:) = F3(x2);

(also tried 'pchip' and 'cubic')

I have to use griddedInterpolant because I need to speed up the whole calculation and because I´m working with a parametrised function. All other functions (like cscvn or csape) take too long. My question is: Is there a possibility to set the first derivative of the first and the last point to 0 or to calculate the natural cubic spline and not a not-a-knot spline?

Thank you for your help!