I had a possibly similar problem calculating probs related to premium bond permutation. With 10^12 memory ran out v quickly. In the end I got round it by writing a recursive function and quantising the probability density function.
Elliot Temple wrote: > Problem: Randomly generate 10 integers from 0-100 inclusive, and sum > them. Do that twice. What is the probability the two sums are 390 apart? > > I have code to do part of it (below), and I know how to write code to do > the rest. The part I have calculates the number of ways the dice can > come out to a given number. The problem is the main loop has 9 > iterations and it takes about 2.5 minutes to begin the 4th one, and each > iteration is about 101 times longer than the previous one. So: > >>>> x = 2.5 * 101**6 >>>> x /= (60*24*365.25) >>>> x > 5045631.5622908585 > > It'd take 5,000 millennia. (If my computer didn't run out of memory > after about 4 minutes, that is.) > > Any suggestions? Either a way to do the same thing much more efficiently > (enough that I can run it) or a different way to solve the problem. > > Code: > > li = range(101) > li2 = [] > range101 = range(101) > for x in xrange(9): > print "x is %s" % x > li2 = [] > for y in li: > for z in range101: > li2 += [y+z] > li = li2 > print li.count(800) > # prints how many ways the dice can add to 800 > > > This link may help: > http://www.math.csusb.edu/faculty/stanton/m262/intro_prob_models/calcprob.html > > > -- Elliot Temple > http://www.curi.us/blog/ > -- http://mail.python.org/mailman/listinfo/python-list
