The Power of Probabilistic Scenarios in Constantly Changing Supply Chains

Heraclitus, the ancient Greek philosopher, captured a fundamental truth about existence: "Change is the only constant in life." – Heraclitus. ?In the world of data modeling, adaptation and evolution are equally paramount. This brings us to the crucial concept of probabilistic scenarios. But what exactly are probabilistic scenarios?

Let's break it down in simple terms.

Probabilistic scenarios are sets of data points generated to represent potential real-world situations. Unlike scenarios in war games or other simulations, these are time series data that serve as inputs for various systems or decision-making processes.

In this blog, we'll cover the process of creating scenarios, targeting both standard and often overlooked areas. We'll explore the sources of scenarios and explore the criteria for evaluating scenario generators. Additionally, we'll walk through how to generate univariate and multivariate scenarios using a time series bootstrap method.

Now, let's talk about how these scenarios come to life. Typically, we encounter scenarios generated by predefined models offering various parametric options. However, there's a notable drawback in these simplified analytical models - they don't always capture the complexity of real-world data.

That's where the concept of bootstraps steps in as an alternative approach. It allows us to work directly with actual real-world data, converting it into scenarios for a more accurate representation. There are two primary applications for bootstrap scenarios. The conventional use involves transforming a single real-world scenario into an unlimited number of artificial scenarios. These can be employed as inputs for simulation models, enabling stress testing and performance estimation for system design.

The second application is often overlooked but equally important. Bootstrap scenarios can serve as inputs for decision-makers or system operators. This visual representation aids in providing valuable insights for critical decision-making processes.

Now, let's discuss the various sources of scenarios. We've identified four key origins:

1.????? Geppetto's Workshop: This involves crafting scenarios manually by experts. While it can yield high fidelity, it tends to be resource-intensive and may not generate a large number of scenarios.

2.????? Groundhog Day: This method involves repeatedly using a single real-world situation as input. While it's cost-effective, it lacks variety and may not accurately reflect the diversity of real-world scenarios.

3.????? Parametric Models: These models offer a larger number of inputs, but they may not always capture the complexity of real-world data, potentially compromising fidelity.

4.????? Non-Parametric Time Series Bootstraps: This approach allows for fidelity, variety, quantity, and cost-effectiveness. It's a versatile method that excels in creating realistic scenarios.

Next, we will focus on the principles for evaluating scenario generators. It’s important to discern the quality and applicability of different scenario generators. To do this, we rely on four fundamental principles: fidelity, variety, quantity, and cost.

Fidelity is about how accurately a scenario imitates real-world situations. High fidelity means it mirrors actual events closely, giving a solid foundation for analysis and decision-making.

Variety covers the diversity of scenarios a generator can create. A versatile generator can simulate a wide range of potential situations, allowing for a thorough exploration of possibilities.

Quantity refers to how many scenarios a generator can produce. A generator that can generate a large number of scenarios provides ample data for analysis.

Cost considerations are crucial in practical applications. An efficient scenario generator balances quality with resource usage, ensuring the investment is justified by the value and accuracy of the outcomes.

Visual Scenarios: Enhancing Understanding and Response:

Now, let's shift our focus to the visual aspect of scenarios. Visual representations play a pivotal role in training and decision-making processes. They offer a tangible and intuitive way to grasp potential real-world situations, enhancing the comprehension and response of operators.

We bridge the gap between abstract numbers and actionable insights by translating data into visual scenarios. These visual aids empower operators to swiftly and accurately interpret information, enabling them to make informed decisions in dynamic, time-sensitive situations.

In essence, visual scenarios serve as a vital tool in our arsenal, ensuring that our models and analyses are not only accurate but also accessible and actionable in the real world. They transform data into a language that can be readily understood and applied, ultimately enhancing the effectiveness of decision-making processes.

In conclusion, understanding and effectively utilizing probabilistic scenarios is instrumental in refining and enhancing our models. We can select the most suitable method for generating scenarios tailored to our specific needs by considering the various sources and criteria.

Remember, while scenarios are powerful tools, they have their limits. They can significantly contribute to system design and analysis, but they should be used within the bounds of their capabilities.

Those interested in further technical details and references can be found here.

What makes a probabilistic forecast?

Probabilistic Forecasting for Intermittent Demand