ONE Record Hackathon Toronto
Air Canada Cargo
Written Report
Challenge 1 – Warehouse
Delivered by IBM Bee Hackers Team:
Kyra Disimino
Maya Michniewicz
Robert (Ben) Wallace
Angela (Yunxuan) Weng
Xinran Xiong
Page Break
1.0 Overview
The following section outlines the challenge that our team is aiming to solve and the scope of our solution.
1.1 Introduction to the Team
Our team’s name is the Bee Hackers. We are made up of 5 IBM Associate Consultants and Developers with backgrounds in computer science, business and engineering.
1.2 Background
In the current warehouse business, there are multiple factors that affect the efficiency of the operational process, such as the spatial layout, inventory accuracy, pick path and so on. Receiving, storage, and shipping are the 3 primary components that comprises a warehouse. Research shows that 55-65% of the warehouse operational cost is allocated towards order picking [1]. In addition, order picking accounts for roughly 40-50% of all activity time and within the order picking activity roughly 50% of the time is spent travelling to and from pick locations [2]. This points to the importance of inventory organization and storage space. Therefore, optimizing the use of dock space for storage and inventory slotting would be critical for achieving optimal pick paths, increasing operational efficiency and reducing overall cost.
1.3 Scope
Our team has decided to take on the Warehouse Challenge, focusing on picking optimization, and working towards improving processes in warehouse operations, including mapping of cargo docking as well as identifying slotting locations with shortest path and greatest efficiency. To do so, we plan on using the ONE Record API to create an innovative mapping tool to optimize warehouse management including use of dock space and inventory slotting for achieving optimal pick paths.
2.0 Solution
The following section explains the team’s solution approach to the Warehouse Challenge.
2.1 What We Built
The team created a database from our own warehousing company, Bee Hackers Ltd., which has 2 isles and 36 slots. To address our chosen warehousing issue, we built an innovative tool that generates reports and warehouse maps in order to anticipate incoming cargo and make data-driven decisions for slotting strategy. Additionally, it will minimize the pick path length to further reduce costs and increase efficiency.
We leveraged the One Record API to perform predictive analytics, in order to optimize pick paths of future orders. We used logistics objects along with the One Record API to create digital twins of warehouse items and locations to simulate real-time results.
2.1.1 Links to Test the Solution
The following is the link to the team’s presentation video: Team Bee Hackers - ONE Record hackathon submission
Github link: https://github.com/rbenwallace/one-record-server-java
Other relevant files (code, Excel, mapping) are zipped in the submission folder.
2.2 Impact of the Solution
As shown in the solution demo linked in the previous section of the report, the main KPI in this case is the average pick per hour, which impacts the operating costs as well as the pick cycle. Using mapping algorithms and visualizations, the team was able to increase efficiency in the pick process by an average of 30% as shown in table 1. With a warehouse averaging an estimated 70 picks per hour, our solution has optimized this to 100 picks per hour.
Table 1. Efficiency calculations in Excel
Assuming the average warehouse worker makes 28.86$/hour, and they work 24/7 and 1 warehouse worker is needed per pick, this leads to an average total cost reduction of approximately 3.89M$ per year.
Assuming the majority of costs associated to order picking in the aforementioned operating costs diagram are related to the labour associated with slower pick times, our solution has effectively reduced the percentage of operating costs related to order picking to 29.4%.
2.3 Limitations
Although improvements are shown in the team’s solution, there exists limitations in the model that includes the simulation of a small warehouse, small number of pickers, and lack of consideration for cargo size variation. The further steps to address these limitations are described in the following section of the report.
3.0 Next Steps
As this solution is extended to a warehouse with fewer aisles, the team’s next steps would be to scale the solution to a larger warehouse. Ideally, the team would be including different data models from the ONE Record API to further optimize picking strategy including but not limited to: dimensions, sensors and stackability in order to improve inventory accuracy and space optimization; Expiry date, handling instructions and others to further improve on the picking strategy.
Page Break
4.0 References
[1] R. Key and A. Dasgupta, “Warehouse Pick Path Optimization Algorithm Analysis,” FCS Papers, 2015. [Online]. Available: http://worldcomp-proceedings.com/proc/p2015/FCS2609.pdf. [Accessed: 23-Oct-2022].
[2] M. Badwi, “Using warehouse automation – optimising man and machine for maximum efficiency?” Supply Chain Junction, 2006. [Online]. Available: https://www.scjunction.com/blog/warehouse-automation-optimising-man-and-machine. [Accessed: 23-Oct-2022].
[3] “Picking orders in a warehouse: Tips, Strategies & Best Practices,” 6 River Systems, 25-Aug-2022. [Online]. Available: https://6river.com/warehouse-order-picking-tips-strategies-and-best-practices/#:~:text=Warehouse%20order%20picking%2C%20in%20particular,multiplied%20when%20automation%20is%20used. [Accessed: 23-Oct-2022].
Log in or sign up for Devpost to join the conversation.