Article ID: | iaor20132308 |
Volume: | 142 |
Issue: | 2 |
Start Page Number: | 290 |
End Page Number: | 301 |
Publication Date: | Apr 2013 |
Journal: | International Journal of Production Economics |
Authors: | Pathumnakul Supachai, Piewthongngam Kullapapruk, Homkhampad Suphakan |
Keywords: | production, agriculture & food, combinatorial optimization, statistics: inference, datamining |
In an integrated feed swine company, the firm engages in weekly planning of feed logistics to ensure the proper growth and health conditions for the pigs, taking into account constraining factors, such as farm feed inventory, inventory policy, and the warehouse capacities of individual farms, as well as other logistical constraints. This process becomes a more complicated task as the number of farms rises. The studied problem was a case of multiple feed mills, multiple farms, and multiple products, which resembles a multi‐facility, multi customer, and multi‐product case. Database management coupled with a mathematical modeling method is proposed to cope with the industrial‐scale feed production–distribution planning and to determine the most suitable feed delivery cycle, number of trucks used, feed order quantities for individual farms, and production batch size of the feed mill such that the cost from mill to feed is minimized. The performance of heuristic solution approaches has been assessed. The solutions from the proposed heuristic algorithm were shown to slightly deviate from the optimal solution but the computational time declined dramatically. In addition to the efficiency gained from the optimization framework, the database‐centric feed demand estimation likewise enabled an accurate feed demand, taking into account pig health and growth conditions, as well as other necessary input factors for the model. The overall decision‐support system works interactively with the optimization model and shows promise when applied to real industrial problems.