我有这一部分在组中制作一些随机链接。我可以提高效率吗?有什么想法吗?
感谢。
n=[10 ;10 ;10];
no_of_groups=size(n,1);
for k=1:no_of_groups
if k==1
m{k,1}=randi(n(1), n(1),2);%random links between nodes of group1
m1=rand(n(1),1)/2+0.5;%random weight to edges
c{k,1} = cat(2,m{1},m1);
else
m{k,1}=randi([((k-1)*n(k-1)+1) (k*n(k))],n(k),2););%random links between nodes of group k
m1=rand(n(k),1)/2+0.5;
c{k,1} = cat(2,m{k,1},m1);
end
end
c=cat(1,c{:});
答案 0 :(得分:1)
快速解决方法是在循环外预分配 m
和c
以及 initiliaze 以删除if条件:< / p>
n = [10 ;10 ;10];
no = size(n,1);
% Preallocate
m = cell(no,1);
c = cell(no,1);
% Initialize m and c
m{1} = randi(n(1), n(1),2);%random links between nodes of group1
m1 = rand(n(1),1)/2+0.5;%random weight to edges
c{1} = cat(2,m{1},m1);
for k = 2:no
m{k} = randi([(k-1)*n(k-1)+1 k*n(k)],n(k),2);%random links between nodes of group k
m1 = rand(n(k),1)/2+0.5;
c{k} = cat(2,m{k,1},m1);
end
c=cat(1,c{:});
这个循环是可矢量化的,或者至少可以避免使用单元格,这样可以提高速度。