Course 1, Week 4

This programming assignment is an oversimplified simulation of the page rank algorithm. It uses matrix multiplications and the principle of the Markov matrix. You can predict the future probability of being redirected to each page if you have the vectors of the current probabilities and the probabilities one step behind. The matrix P will be the only unknown in the equation that describes the eigenvalue of the Markov Matrix P (i.e. 1), X_{m} and X_{m-1}.

When I give all pages an equal probability of being linked to in P, the Exercise 4 check fails. The multiplication result is different from the Eigenvector. Despite that, the auto-grader still assigned me the same score as when I assigned different and random ranks to the pages. Why is the multiplication result different from the eigenvector just because all pages have equal probability according to P? I swap the two versions of P and the exercise fails on the equal probabilities but passes the random ones.