# Perceptron weight vector update

I read about the Rosenblatt Perceptron Learning Algorithm. Often there is an explicit note:

It is important to note that all weights in the weight vector are being updated simultaneously

But why are all weights updated simultaneously? I tried another approach where I iterated over all weights and updated them in different iterations. It also worked on some simple test cases.

Could someone explain to me, why they are updated simultaneously and why should this approach be better?

• here is proof of convergence (pp.15) where they call the margin $\delta$ – oW_ Aug 4 '17 at 16:22