Abstract:

The character of usingsymbols to express the logic relation in the GDP model was fully considered. A new logicbased branch and bound algorithm was proposed. First, some Boolean variables were determined by using technique character and the principle of choosing branch. Then branch strategies were used to determine the relationship between logic variables. Finally, other Boolean variables were determined by using logic reasoning. A typical batch process productionscheduling model was solved using detail steps of the algorithm, and the results were analyzedand explained. The simulation results show that solution efficiency can be improved because a lot of branches are pruned through logic branches and reasonable use of the relations among the Boolean variables in the model, and the optimum can be obtained after a few nodes are examined.

Key words: production scheduling; generalized disjunctive programming; logic branch; algorithm