We encourage you to open the file 'demo.jl' and modify it to try out different precision of the approximation and different systems. Whenever you have changed the parameters to your liking, simply run 'include('demo.jl')' in the Julia REPL again to see the new result.
We encourage you to open the file `demo.jl` and modify it to try out different
precision of the approximation and different systems. Whenever you have changed
the parameters to your liking, simply run `include("demo.jl")` in the Julia REPL
again to see the new result.
## Changing Script Behaviour
You can specify if the script should run a validation, if it should search for
the smallest feasible shape parameter, and if it should generate plots by changing
the following boolean variables in `demo.jl`:
```Julia
validation = true
plotting = true
search_c = true
```
## Changing the System
You can change the system matrix $`A`$, the (diagonal) quadratic cost weight
matrix $`Q`$ and the total cost $`J`$ by changing the following variables in
`demo.jl`:
```Julia
A = [0. 15.;15. 0.]
Q = diagm([1., 1.])
J = 1.
```
Note that `A` and `Q` should have elements of type `Float64`, while `J`
should be of type `Float64`. This is most easily achieved by adding a decimal
point after each number, as done above.
## Changing the RBF Approximation
You can change the number of collocation points in each dimension, the value of
the shape parameter and the boundary values of the grid for the collocation points
by changing the following variables in `demo.jl`:
```Julia
xlimits = [[-1.5, 1.5], [-1.5, 1.5]]
Ni = [30, 30]
c = 80.0
```
Note that `xlimits` should be of the type `Array{Array{Float64,1},1}`, i.e. a
vector of vectors, where each element contains the lower and upper bounds of the
grid for each dimension. `Ni`, which specifies the number of collocation points
in each dimension, should only contain elements of type `Int64`. The shape
parameter `c` should be of type `Float64`. Note that if `search_c=true` has been
specified, then the value of `c` will only be used as an initial value in the binary search
