Supply Chain Mgmt. Simulation in the Beer Pub

By Ron Sengupta


Journey Begins- Discovering the World of Supply Chain


I have recently signed up for Supply Chain Management Micromasters at MITx. This is a fascinating and rich program that dive deep into supply chain systems and technology. The program introduces four core components of supply chain technology, 1. Data Management, 2. Machine Learning for SCM, 3. Supply Chain Systems and 4. End to end simulation from the perspective of the supply chain management ecosystem.


The simulation aspect of the supply chain is completely new to me. Although the formal program has yet to touch on simulation, it caught my attention (kind of a shiny new thing). So, I started exploring it on my own. This is my musing.

 

The Role of Simulation in Supply Chain Management


As per  the “Supply Chain Management: Strategy, Planning, and Operation“ by Chopra and Meindl “A supply chain consists of all stages involved, directly or indirectly, in fulfilling a customer request. The supply chain not only includes the manufacturer and suppliers, but also transporters, warehouses, retailers, and customers themselves.”


The primary stages of supply chain management are planning, sourcing, production, distribution, and returns. 



Planning: This is the stage where companies strategise on how to manage the resources to meet customer demand efficiently. It involves forecasting, demand planning, and establishing the supply chain structure. Simulation is key in this stage because it allows companies to model different scenarios, predict outcomes, and optimize their strategies before implementing them. This helps in making informed decisions about capacity planning, inventory management, and logistics.


Sourcing: This stage involves selecting suppliers and managing supplier relationships to procure raw materials and components.  Simulation might be able to play a role here, but less critical compared to the planning stage.


Production: This involves transforming raw materials into finished products. Simulation can be used to optimise manufacturing processes, improve production schedules, and ensure efficient use of resources.


Distribution: This stage involves delivering the finished products to customers. Simulation can help optimise transportation routes, manage warehouse operations, and improve delivery performance.


Returns: This stage deals with the return of products from customers, either for recycling, or disposal. Simulation can be used to streamline the returns process and manage reverse logistics efficiently.



In conclusion, while simulation can be applied in various stages of supply chain management, its primary and most impactful role is in the planning stage, where it helps in scenario analysis, and optimization of the supply chain network.


Exploring Types of Simulation


Here are a two common types of simulations,


Discrete Event Simulation (DES)

“DES models the  operation of a system as a sequence of events in time.” Think about simulating a pub, where events such as customer arrivals, drink orders, receiving drinks, occasional fist fights and finally leaving the pub  occur at specific times.In this scenario, the simulation tracks each event discreetly. The system state only changes at these discrete points.


Image Courtesy: Wikipedia




Continuous Simulation 

Continuous Simulation models systems where changes occur continuously over a period. It typically uses a differential equation. Think about  simulating the operation of a commercial coffee machine, the simulation would continuously monitor and adjust variables like water temperature, coffee flow rate, and milk levels throughout the day. Unlike discrete simulations where changes occur at distinct events, here changes are ongoing and the state of the coffee machine (like temperature and pressure) is continually updated in real time as customers use the machine.


There are other types of simulations as well (like agent based simulation)  but currently I am not exploring them.  

DES is extensively used for logistics and operation. It involves models to replicate the behavior and interaction of components in a supply chain. The purpose is to study how changes in the system affect overall performance, such as the impact of changes in warehouse operations, or logistics. This can help in understanding bottlenecks, validating theories about improvements, and making decisions.


Zooming In  @DES- The Beer Pub Experiment 


Welcome to "The Thirsty Developers" , a beer pub somewhere near Holborn. It's the kind of place where everyone seems to know your name, even if they don't.


The Characters


Bartender-  Sammy, he is  a magician with a beer tap and he somehow knows when you need a refill. Unfortunately we can’t clone him yet, so he can't be in two places at once. We tried cloning him, it did not work.


Customers- Mostly after-workers and each comes with their unique thirst level and patience meter. 



The Challenge?


To manage Sammy’s time and attention and to keep all glasses at least half full (because we're optimists here). Just when you think it’s all smooth sailing, a bus full of thirsty tourists might enter the pub.  How long will each customer wait? Can Sammy keep up with the rush or will he need to call in reinforcements? Only our simulation can tell.


Using the power of SimPy (It’s an open source python library primarily used for discrete simulation), we simulate, so grab a pint.

 

The goal was to model a typical evening at a pub with a focus on customer service processes. Here’s a explanation of the thought process [ Please note, I have used it  for learning and so the model is quite simple]


Thought Process


Customers and Bartenders: Identified as key players. Customers follow their own schedule: arrive, order, and drink. Bartenders are the critical resource needed for orders to be processed.


Staggered Arrivals: To mimic real-world scenarios, customers don't all arrive at once. Instead, they come in at different times, selected randomly from a predefined list, to simulate the natural flow of a pub.


Simple

Kept customer actions straightforward to focus on the impact of bartender availability on service times and customer wait times.


Using SimPy, the simulation tracks discrete events such as arrivals and departures.


Resource: Bartenders are managed as limited resources, affecting how quickly customers are served.



Operational Insights: The simulation throws light on potential bottlenecks and can help explore strategies to improve CSAT.



Analysing the output from Beer Pub simulation using the 80-20 Pareto principle, 


Customer Wait Times

  •    There are considerable wait times between when customers arrive and when they receive their orders (e.g., Customer3 waits 18 minutes, Customer8 waits 38 minutes).

  •    Waiting time  increases for customers who arrive later, indicating a buildup of demand that Sammy struggles to meet.


Inefficiencies could be attributed to

 

1] Insufficient Bartender Resources

 

 2] Arrival Patterns - The clustering of customer arrivals exacerbates the issue, as the bartender is overwhelmed during these peak times.


Recommendations for Improvement


1. Increase number of bartender


2. Implement a Queue Management System [ Incentive for free beer may work]

   


Code

https://github.com/ronsengupta/randomcode/tree/master

Comments

Post a Comment

Popular posts from this blog

Baseline(s).. [Part 1]

Image
Merriam-webster defines baseline "as a line serving as a basis", or you can think of it is a starting point used for comparisons. In site reliability engineering (SRE), establishing a performance baseline for workloads is a common practice. This baseline serves as a reference point for monitoring and also improving system performance over time. In my previous experience as a SRE, particularly in supporting critical e-commerce websites, I learned that it's beneficial to establish more than one baseline for a more holistic view and accurate analysis. I have found two key types of baselines are quite useful: [A] Performance Baseline Influenced by External Factors: This baseline takes into account real-world conditions such as network variability, user device performance, and other external influences. It reflects the performance metrics as experienced by the end-users, providing insights into how these factors impact the user experience. [B] Performance Baseline Detached ...
Read more

Art of unsettled science

               It's science but...not all variables are "settled"... Welcome to "Art of Unsettled Science" – our little corner of the "real world computing" we talk about topics in the realm of distributed systems that are little more than a textbook full of facts. Here, we dive into these grey areas that are not quite "settled" yet, where theories are still in the making, and everything feels a bit like a work in progress. We are all about exploring the unknown with a sense of wonder and a healthy dose of curiosity. So, if you're into adventures where the answers aren't clear yet, you've found your crew. 
Read more