| Summary: The aim of this study was to apply aggregate production planning in a fiber board manufacturing factory as an example of the forest products industry. Sales and storing decisions of types, periods and amount of products within the constraints were analyzed by using the linear programming (LP) model of the dynamic state. Before setting the LP model, in planned periods, decisions on the products produced were made by Pareto analysis. The artificial neural network (ANN) method was used to forecast of sales for future periods after reaching decisions on the products. The objective function of linear programming consisting of 15 product types and 180 decision values was obtained for optimum product composition with maximum income by using 207 constrains and 65 iterations. In order for production management to make better decisions, sensitivity analysis was applied to obtain the best LP model. When the obtained model was used, the unused capacity of the press line was found to be 15.5% regarding all production periods according to sensitivity analysis. The results showed that the factory could produce an average of 9,453 m3 more products i.e. 60,971 m3 productions instead of 51,518 m3 for four terms if the proposed model was used. When production time was considered, it was observed that 18.77 hours of working time would be sufficient for the same production amount, instead of 22.22 hours. It was concluded that successuflly using the decision making techniques such as LP reduce the production losses and provide a huge contribution on productivity and profitability on Forest Products Industry.
Key Words: Production planning, Linear programming, Artificial neural network, Forest products industry |
