Random search in the one-dimensional cutting stock problem

The paper reports on a randomized approach to the one-dimensional cutting stock problem. In a sequential heuristic the next pattern to be processed is chosen according to the best outcome of 200 trials of random bin packing. The quality of the solutions is discussed and the problem of multiple occurrence of equivalent solutions with different patterns is explained by the cross over technique known from genetic algorithms. For one data instance with 27 demand lengths, reported by Haessler (1975), four different optimal solutions are given.