Study On Modeling And Optimization Algorithms For Steel Tube Order-grouping Problem In Iron And Steel Industry

Steel tube order-grouping, which determines the corporate earnings and evaluates thecompetence of the company, is one of the most important problems in steel tube enterprise.The planner organizes multiple orders with the same specifications producing on one hotsteel tube, of which the objective is minimizing surplus. With the sharp market competition,the production requirements are becoming small-volume and high-variety. Responding tothe changing requirement and reducing the production cost have become the key issues tothe development of steel tube enterprise. Therefore, studying steel tube order-groupingproblem is of great realistic significance.According to the processing technic of steel tube, there are two types of steel tubeorder-grouping problem: First, enterprises buy steel tubes, of which the hot cutting is notexist; Second, steel tube enterprise use of steel tubes their own, of which the effect of hotcutting has to be concerned. The second type is investigated in this paper. It is a typicalcombinatorial optimization problem that widely spreads in the steel tube industry. Thispaper studies steel tube order-grouping problem aiming to meet the need on site.With the characters of steel tube order-grouping problem, a nonlinear mix-integerprogramming model is formulated. The object is minimizing surplus and the mainconstraints including equipment specification limits, the total order length and unit lengthare both interval values, hot cutting length requirements etc. are all taken intoconsideration in the model. For the sake of solving this problem, the programming modelis transformed into a linear mix-integer programming one.Due to the large scale and real-time on site, a hybrid method combined with heuristic andMILP is proposed in this paper. There are two phases in this method: The one is dealingwith the main part of the orders, using a mix-integer programming model to deal with thelow flexible orders firstly, and then relaxing some constraints to deal with the otherunfinished orders. The other is to finish the order-grouping of the rest orders with the helpof surplus and mix-integer programming model. This method makes use of the advantageof both heuristic and mix-integer linear programming. It decomposes the problems into afew small ones and optimizes the objective step-by-step, which is flexibility and rapidity.At the same time, the use of mix-integer linear programming assures the accuracy of the solution.Computational experiments with practical data demonstrate that this hybridmethod is able to find a near optimal solution within an acceptable time, which fully meetsthe need of site.Based on the hybrid method, steel tube order-grouping system is developed. The system isable to give the result quite fast and effectively reduce the surplus, which avoids theshortcoming of the planners order-grouping and provides strong decision support forplanners.