Even if we choose the cluster centers initially still there is no guarantee that we will get the same clusters every time.
There is one paperon this and a specific clustering technique to solve this problem But I am unable to code it properly. Can you have a look on this algorithm called Quality Threshold (QT) based clustering.
Heyer, L. J., Kruglyak, S., Yooseph, S. (1999). Exploring expression
% data: Identification and analysis of coexpressed genes. Genome Research
% 9, 1106–1115.