Browsing by Author "Ertem, M.A."

Now showing 1 - 2 of 2

Citation - Scopus: 0
Forecasting day of week volume fluctuations in the intermodal freight transportation
(Institute of Industrial Engineers, 2011) Ertem, M.A.; Ertem, Mustafa Alp; 52460; Endüstri Mühendisliği
Average daily volume fluctuates intensely based on the day of week in the intermodal freight transportation. Shippers tend to peak around Thursdays and receivers tend to peak around Mondays. These fluctuations bring challenges to the industry in terms of capacity management and getting reliable service from the railroad companies. The purpose of this study is to forecast J. B. Hunt Transport Services, Inc.'s load volume on railroads. Load is meant to be the number of containers that will arrive at a rail ramp during a 24hrs time window. The end in mind is to have better service from the railroad companies and to manage the company owned equipment better. The forecasting model applied to tackle this problem is a multiple linear regression model and is based on the historical in-gate numbers. It uses the previous two year's data and day of week information as independent variables, and current year's data as the response variable. The results indicate better accuracy levels for the model when compared to the two week moving average.
Citation - Scopus: 22
Using announcement options in the bid construction phase for disaster relief procurement
(2012) Ertem, M.A.; Buyurgan, N.; Pohl, E.A.; 52460
This paper presents an analysis of the bid construction phase of procurement auctions in disaster relief and humanitarian logistics. Substitution and partial fulfillment options are presented in formulations to allow bidders with fewer inventories to offer substitute item types and partial bids in auctions. During the auction announcement phase, a coordinating platform for disaster locations (i.e., auctioneer) allows substitution and partial fulfillment options to the relief suppliers (i.e., bidders) when acceptable. Thus, suppliers with fewer inventories can offer substitute item types and participate in more auctions by partially bidding. A genetic algorithm, a simulated annealing algorithm and an integer program are used for the analysis of the bid construction phase with different announcement options. Heuristic solution techniques and an IP formulation help understand the dynamics of the bid construction problem. It is shown that the addition of substitution and partial fulfillment options is essential to diversify and increase the usable capacity of the supplier base. Additionally, the partial fulfillment option enables better usage of supplier inventories in an environment with scarce supplies. © 2012 Elsevier Ltd.