In progress Sustainable Logistics in Fresh Food (SLIFF)

Reference:
ERIM PhD 2014 RSM NWO RdK_MB

Abstract

There is a growing global need to improve the distribution and merchandising of fresh foods by maximizing fresh shelf life while minimizing supply chain costs as well as environmental impacts. The central objective of this project is to enhance the sustainability of fresh food logistics by improving logistics resource utilization while distribution lead times are reduced. In particular, the PhD project focusses at the optimization of the operations within a fresh food warehouse network. Decisions need to be made regarding the joint optimization of transportation, inventory levels and material handling resources between primary and secondary distribution centres as well as retail locations.

This PhD project is part of a larger collaboration between the Erasmus University Rotterdam, University of Groningen and a partner retailer with a contracted logistics service provider (LSP) who, in the future, could act as a cross chain control centre to coordinate transportation and inventory optimization for the many vendors involved.

Keywords

Facility and distribution logistics, transportation, inventory control

Time frame

2015 - 2019

Topic

Currently, fresh food is largely cross-docked by retail chains or stored for a brief period of time in the retailer’s distribution centres (De Koster and Neuteboom, 2001). In this PhD project we mainly focus on two problems related to transporting and handling these fresh food flows:

  1. coordination of truck scheduling decisions to arrange the internal operations in the distribution centre, and
  2. coordination of space usage within the DC for fresh food vendors.

Our objective is to develop optimization models for both problems. We first elaborate on both problems separately.

Many retailers operate multiple warehouse facilities, often with different functions. Slow moving fresh and ambient products are picked by store in national warehouses and from there shipped to regional warehouses, where the goods are cross-docked and merged with other flows destined for the same stores. The truck inbound and outbound schedule at the regional distribution centre is determined by taking into account desired delivery time windows at the stores, dock door availability at the regional distribution centre, and projected time needed at the dock doors to handle additional loads that need to be merged. Literature on transport-cross-docking operations usually neglects the internal processes. That is, it focuses on truck-to-dock-door assignment or on truck scheduling at the dock doors (Van Belle et al., 2012; Boysen and Fliedner, 2010; Bodnar and De Koster, 2013), sometimes as part of a vehicle routing problem (Wen et al., 2009). Such literature assumes arrival and departure time windows are given and internal processes in the distribution centre have ample capacity. However, if additionally products have to be retrieved from storage, the order picking schedule must match the docking schedule. For instance, if a large quantity must be retrieved from inventory for a retail store, this will impact the time that needs to be reserved at the outbound dock lane for that store. This implies that not only the cross dock quantity and timing influences the outbound time schedule, but also the order pick schedule. In this research we want to include the internal processes in the distribution centre in the decision-making process and thereby make cross dock schedule models more realistic. Such inclusion implies that incoming loads may have to be buffered temporarily, that freight flows from the storage function inside the DC have to be merged with the cross-dock flow, and that the time needed at the dock doors will be influenced by the quantities that have to be cross docked and picked from inventory. The main research questions we want to answer are:

  • What is the impact of internal storage and handling processes on the truck schedule for a cross dock process at a distribution centre?
  • How can internal processes and cross docking be integrated in the truck schedule at a distribution centre with the objective to minimize the total makespan of the process or to minimize the number of doors needed to handle all flows?

The chosen objectives in the second research question are important, as they allow to increase capacity of the cross dock without physical expansion in handling capacity or space capacity. Consequently, handling a larger number of trucks at a smaller distribution centre will lead to lower emission levels, since warehouses contribute to about 9% of all carbon dioxide emissions.

The second topic of this research project focuses on the management of inventory space at the distribution centre in relation to the frequency of inbound transport for inventory replenishment. Such space is often managed using vendor managed inventory (VMI) schemes, which implies that fresh food vendors have a limited amount of space in the warehouse at their disposal within which they have to guarantee the service level for the retailer. The amount of space is usually small, forcing the vendors to replenish frequently, often more than once a day, with partial truck loads (De Koster and Neuteboom, 2001). As such transport is rather inefficient, cost savings can be achieved, if vendors manage the space jointly and also organize the transport jointly. This will also substantially reduce CO2 and PM10 emissions. Although in many cases it may be hard to achieve such cooperation, in a distribution warehouse where the retailer owns the space, he can enforce space cooperation and even transport cooperation, e.g. the retailer or an LSP collects the products at the vendors’ warehouses. Integrating the available space over multiple vendors leads to a joint replenishment problem, albeit with individual service level requirements. Joint replenishment problems have been studied in literature (see e.g., Khouja and Goyal, 2008). We integrate this problem with joint transport organization, e.g. by collection of the products at the vendor warehouses by the retailer. This second problem is called ‘factory gate pricing’ and can lead to substantial transport and emission savings, compared to deliveries by individual vendors (see Le Blanc et al. 2006). In this research we integrate the joint replenishment and the joint transport problem given space restrictions and minimum service levels for products of individual vendors, with the objective of minimizing total relevant inventory and transport costs. Our research question is therefore:

  • What is the cost and service impact of jointly managing inventory and transport decisions by independent vendors supplying a retailer distribution centre with a common space restriction?

Our objective is to develop optimization models for joint ordering and replenishment that trade off cost and service objectives as well as CO2 and PM10 emissions.

Supervisory Team

René de Koster
Professor of Logistics and Operations Management
  • Promotor
Marco Bijvank
Marco Bijvank
  • Daily Supervisor