0 个投票

Hello,
I am currently fighting with (more than) double-precision numerics, with which I am not familiar, so the answer probably lies within this topic. I would like to calculate
with and but obviously the precision of will be bad for big k. Is there a datatype of which I am not aware, or any other trick that will help me? Thanks in advance. :)

请先登录，再回答此问题。

 采纳的回答

Rik
Rik 2019-10-14

0 个投票

It will probably be helpfull to use the Symbolic Math Toolbox, and use the tips explained here. For mod(a^b,k) there is a direct function, but for your use case there doesn't seem to exist one.

1 个评论
显示 -1更早的评论 隐藏 -1更早的评论

Eric Schöneberg
Eric Schöneberg 2019-10-15
I've generated the following function:
function A = matrixPower(Matrix, exponential, precision)
A = vpa(eye(size(Matrix)), precision);
for k=1:exponential
A = vpa(A*Matrix, precision);
end
A = double(A);
end
which seems to work. I've tested the method using this script:
A = magic(5);
precision = 100;
for k = 0:50
B = matrixPower(A, k, precision);
C = A^k;
Error = B-C
end
and the Error(-Matrix) was Zero for k=1:10 and grew to 1.0e+74*Matrix for k = 50. Thanks again for your input @Rik.

请先登录，再进行评论。

更多回答（0 个）

类别

在 帮助中心 和 File Exchange 中查找有关 Calculus 的更多信息

产品

版本

R2019a

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by