Abstract
In the current and future Supply Chain landscape, we need to ensure that we keep the warehouse ship sailing amidst the turbulent waters of dynamic business growth and rapid changes in technology. There are several challenges to be overcome, and many opportunities to be embraced on the path to achieving this. In this chapter, the authors detail one of the key problems facing the warehouse and that is Slotting. They look at the various business drivers, and technological drivers impacting Slotting. They propose a solution to tackle this problem by using Market Basket Analysis and Machine Learning.
TopIntroduction
A supply chain encompasses the physical and information flows involved in the buying, manufacturing, and delivering a product or service. Supply Chain Management (SCM) deals with the strategies for sourcing, storing, and delivering the right products to the right locations at the right time to the right markets. Warehouses play a critical role in the supply chain. They help to store the goods closer to the market so that the demands can be fulfilled on time. The major activities in warehousing are:
- •
Receiving,
- •
Storage,
- •
Picking,
- •
Shipping.
Slotting refers to the optimal placement of the items in warehouse locations such that they can be picked in the least time with maximum accuracy and maximum space utilization. Within a warehouse, it is critical to store the right product in the right location for speedy retrieval. This improves the overall efficiency of the warehouse. The key performance indicators for a warehouse (pick efficiency, pick accuracy, and space utilization) can be considerably improved by efficient slotting.
Pick efficiency refers to the speed at which orders can be picked from the storage location and delivered to the shipping location. Pick accuracy measures the percentage of successful picks. Space utilization measures the percentage of wastage in space.
Picking incurs the maximum cost in a warehouse and efficient slotting can significantly reduce warehouse expenditures and increase warehouse efficiencies. Figure 1 illustrates the need for slotting.
From Figure 1, we can see that when products are distributed across aisles, the travel time increases, and correspondingly the picking efficiency decreases. For example, consider an order of 12 units located in 12 locations. The travel time nearly doubles if these locations are spread in four aisles instead of two aisles.
Ensuring efficient slotting for a medium to large scale warehouse necessitates a slotting planning and optimization software solution. A slotting tool is a supply chain execution software application that uses algorithms to create an optimized slotting plan based on a variety of factors (Trebilcock, 2011). It uses variants of slotting optimization software. Till date, various solutions have been deployed for generating optimized slotting plans. However, these solutions are unable to scale to meet the dynamic nature of the present day business scenarios (SCDigest, 2008).
The following business drivers make the existing slotting solutions outdated and shape the need for faster, more flexible and intelligent slotting solutions:
- •
Online retail has increased the scale of businesses manifold and, hence, this requires enhanced slotting capabilities in real time.
- •
Customers demand faster delivery of products with high quality.
- •
Omni-channel commerce has changed the structure and semantics of the supply chain domain. Since this is a dynamic and continuous change, the slotting tool must be agile in nature.
The above business drivers pose increased scalability, speed, and agility demands on the slotting solutions. To the knowledge of the authors, based on their analysis and research, the present day slotting solutions are limited to heuristical or basic mathematical techniques. These solutions do not incorporate state-of-the-art techniques like Market Basket Analysis (MBA) and Machine Learning. The authors propose a new solution which uses these techniques thereby mitigatinag the risks posed by the changing business contour. The usage of MBA and Machine Learning Techniques will help increase pick efficiency, pick accuracy, and space utilization for the warehouse. Efficient slotting will also result in improved direct labor productivity, reduced order shorts, improved ergonomics and ultimately, reduced costs and improved customer satisfaction (KOM International, 2011).
Key Terms in this Chapter
Decision Tree Induction: Decision tree learning utilizes a decision tree as a predictive model and is one of the predictive modeling approaches used in data mining, statistics, and machine learning. The decision tree maps observations about an item to inferences about the item's target value.
Simulation: Simulation is the reproduction of the actions of a real-world process or system over time. The act of simulating something first requires that a model be established; this model represents the key characteristics or behaviors/functions of the chosen physical or abstract system or process. The model signifies the system itself, whereas the simulation signifies the operation of the system over time.
Slotting: The process of placing the right product in the right location in the warehouse with the objective of improving pick efficiency, pick accuracy, and space utilization, and reducing travel time and material handling costs.
Omni-Channel Commerce: Refers to the recent trends in retail commerce where customers demand a seamless buying experience across all channels whether it is online or store-based.
Warehouse Management: Warehouse Management refers to the science of optimizing the flow of goods and information within a warehouse to increase the efficiencies of the various processes and minimize associated costs.
Supervised Learning: This type of learning is similar to the learning demonstrated by human beings, i.e., gaining understanding from past experiences to acquire new knowledge in order to improve the ability to perform real-world tasks. However, machine learning learns from data, since computers do not have “experiences”, which are collected in the past and represent past experiences in some real-world applications.
Market Basket Analysis: Market Basket Analysis is a modeling technique based upon the theory that if an individual purchases a certain group of items, he/she is more (or less) likely to buy another group of items.
Warehouse Management System: Software that automates the key processes in the warehouse from Receiving through Inventory Control to Shipping. It can vary from basic systems to systems performing advanced optimizations.
3D Printing: Technology which utilizes plastic mold, bio-cells, and various other materials to create three dimensional objects.
Supply Chain Management: Supply chain management (SCM) refers the management of the physical and information flow of materials and finished goods in a supply chain.
Machine Learning: Machine learning explores the building and analysis of algorithms that can make predictions from the study of data. These algorithms function by creating a model from sample inputs to generate data-driven forecasts or conclusions instead of being limited to precise fixed program directives.
Internet of Things: The Internet of Things (IoT) is a scenario in which individuals, objects, or animals are provided with unique identifiers, such as IP addresses, and the ability to transfer data over a network without requiring human-to-human or human-to-computer interaction.
Association Rules: Many standard algorithms are available to apply machine learning concepts to practical problems. One such class is association rule algorithms. Association rule algorithms determine the associations between the features used to describe a data set. The origin of association rules mining was as a technique for finding interesting rules from transactional databases.