🚀 go-pugleaf

Can anyone point to nonsequential tests for detecting
changes in mean? I have a set of samples y, where

         y_i = x_i + e_i  i=1...n

and e_i are iid zero-mean normal (variance may be
known or unknown). I assume that x_i is piecewise
constant. Each piece is >=k in length and the jumps
are >=d in height (both d and k known).

I don't know the number of jumps or their locations.
I want to use all the samples at once (nonsequential)
in my test.

Of course, there is a tradeoff in minimizing the
false alarm/misdetection of detecting a jump and
the error in detecting the location of the jump.
I would like a test that takes this tradeoff into
consideration.

Thanks,

Ken Thornton