Date on Master's Thesis/Doctoral Dissertation


Document Type

Doctoral Dissertation

Degree Name

Ph. D.


Industrial Engineering

Degree Program

Industrial Engineering, PhD

Committee Chair

DePuy, Gail

Committee Co-Chair (if applicable)

Bae, Ki-Hwan

Committee Member

Bae, Ki-Hwan

Committee Member

Evans, Gerald

Committee Member

Willing, Gerold

Author's Keywords

order picking; multi-objective; heuristic; optimization; rule mining; expiration date


Order picking is the process of collecting goods and items in specified quantities from storage locations, in response to customer orders. Since many labor resources are involved in this process, finding ways to make it more efficient have been a primary goal for researchers and practitioners. Determining a better allocation of products to the storage areas, finding the best route and sequence to pick multiple products, and choosing the best picking policies to minimize congestion in the aisles are just a few of many objectives regarding order picking process. Due to regulatory compliances and the chance of product spoilage, additional criteria should be taken into account when dealing with order picking in a healthcare warehouse. Following the first-expired, first-out (FEFO) principle and fulfilling an order from a single manufacturing batch, are two of the requirements that are considered in this research. Moreover, minimizing the traveled distance to picking locations, minimizing the penalty or cost of using the space by depleting it quicker, and maximizing the probability of a successful pick in the future by considering the order size distribution are the other objectives that have been studied in this research. The desired order picking problem is formulated and solved as a multi-objective mixed integer programming optimization model. Assigning importance weights to each objective and obtaining one single solution does not provide a full picture of all possible and potentially attractive solutions. On the other hand, providing all the solutions is not always achievable as it is computationally expensive and most of the time a set of rules and regulations drive decision makers toward the chosen solution. Finally, this research focuses on solution simplicity, generality, and practicality since the solution will be implemented by order pickers. To achieve this goal, a novel approach using association rule mining is presented. Products are classified in different groups and some order picking rules are derived based on the relations of these classes together. These rules are then compared with the results of multi-objective optimization model to evaluate their quality. The comparison results showed that surprisingly, some simple rules extracted from the preferences of the decision maker, can obtain good quality solutions. For example, in products with high order variability, it is possible to generate solutions within an average gap of less than ±8.8% from the multi-objective optimization solutions. The findings of this research leads to the following primary recommendations:

  • Orders should be picked based on the expiration dates of the batches (FEFO principle) unless, the amount of inventory in the location is exactly equal to the order amount. In that case, regardless of the expiration date and the traveled distance to the location, it is better to pick from that location and make the space available.
  • If product replenishment is not instantaneous, regardless of the expiration date and the traveled distance to the location, it is recommended to not pick from a location which can lead to an undesired and low amount of inventory. Due to order variability of the product, it is better to wait until the inventory in that location transfers to a desirable level.
  • If fast replenishment of products is obtainable, then prioritizing traveled distance over the objectives linked to space usage and inventory is advisable.