Abstract: The proposed approach solves two-dimensional cutting stock problem using three-block patterns and a sequential value correction framework. Cutting patterns in a cutting plan are generated sequentially. Each pattern is used to fulfill some portion of the remaining items until all the required items are fulfilled. The item values are adjusted dynamically to make the value more reasonable. Multiple cutting plans are generated in order to select the optimal one. The three-block patterns are generated by an incomplete enumeration method, in which each pattern contains at most three types of items to simplify the cutting process. The comparison with the linear programming algorithm indicates that the proposed approach can effectively reduce the plate consumption when the item demands are small.

Key words: two-dimensional cutting stock problem, sequential heuristic procedure, sequential value correction, three-block patterns