Upper bounds for the 0-1 stochastic knapsack problem and a B&B algorithm