Steven_Huang

115 Reputation

6 Badges

4 years, 6 days

MaplePrimes Activity


These are questions asked by Steven_Huang

I notice that Maple transactions journal does not publish any new issue ever since March this year. Based on the previous recording it is supposed to publish issue each season. When is it going to promote a new issue? My research is supposed to publish there and I believe I have made some interesting progress and it is of interest to the community.

So the problem I encountered is when I try to assign ranking of variables of ODE in ThomasDecomposition, the indexed variables are not allowed. For instance, for an ODE system containing a[0], a[1], a[2] as dependent variables, if I assign:
R := Ranking([x],[a[0],a[1],a[2]])
It throws an error:
This is going to be inconvenient when I have to dynamically solve ODE system within a function call, especially when the system is nonlinear and algebraically closed(no integration constants in solutions) and calling DifferentialThomas in dsolve will be slow. I want to know how to resolve this.

When I use Maple to solve some complex polynomial system and trying different methods, for instance, using engine = traditional, it will sometimes lose or fail to find the solutions. In the document, it says Maple will convert to engine = groebner when it does not find solutions using traditional engine. As far as I know it is using resultant to eliminate variables, and the resultant is computed by pseudo remainder. The leading coefficient of the divisor is multiplied during each division to ensure that we include the case that it is zero.

But why this procedure will lose solutions sometimes? What makes Maple abandon this method?

I intend to use LinearAlgebra package to do some calculations. I want to compute the basis for large Matrices. My discovery is that the linalg[kernel] command, which the document claims is deprecated, could do such computation significantly faster than the LinearAlgebra[NullSpace] command. For a 200 x 500 large random matrix, linalg[kernel] clocked 33 secs, while the LinearAlgebra[NullSpace] takes 200 secs, as shown in the worksheet NullSpace_vs_kernel.mw.

I wanna know what makes the difference, or is there a misuse for LinearAlgebra[NullSpace].

 Based on the type of coefficients in the linear equations, SolveTools[Linear] provide several method including method = Rational, Polynomial etc. 
The Polynomial method of SolveTools[Linear] however, cannot be directly called by SolveTools[Linear]. The interpreter complains that no such methd called "Polynomial".

This could only be solved if I import the SolveTools package in advance and call Linear.

I hope Maple could solve this issue. 

1 2 Page 1 of 2