Understanding Uncertainties in System Engineering
Imagine we’re going to bake cookies, and there are two types of uncertainties we might face: epistemic and aleatory.
1. Epistemic Uncertainty (Lack of Knowledge)
Epistemic uncertainty is like not knowing what kind of cookies people will like. Imagine we don’t know if people will prefer chocolate chip or oatmeal raisin cookies. This uncertainty is because we don’t have enough information about people’s preferences.
In system engineering, it’s like not knowing exactly how a new machine will work or what people will think about it in the future. We can do research, ask people, and gather more information to reduce this kind of uncertainty.
2. Aleatory Uncertainty (Randomness)
Aleatory uncertainty is like the randomness in baking cookies. Even if we follow the recipe perfectly, each batch of cookies might turn out slightly different. Some cookies might be a little crunchier, some a little softer, and that’s just because of small random differences in the baking process.
In system engineering, this is the kind of uncertainty that comes from natural variations and randomness. No matter how much we try, we can’t completely get rid of this randomness.
Apple’s Design Example for Epistemic uncertainty
Apple’s design process is a great example of managing epistemic uncertainty without direct surveys. They often rely on other methods to gather information and reduce uncertainty. Let’s break down how Apple might handle epistemic uncertainty in this context:
Apple is designing a new iPhone feature, such as a new user interface (UI) design. Apple’s design process for a new user interface (UI) involves several stages to reduce epistemic uncertainty. Initially, their experienced designers and engineers make design decisions based on their expertise, though there’s uncertainty about user reactions since no feedback has been collected. They then create prototypes and test them internally with employees and focus groups, which helps identify major issues and provides initial feedback. Next, they apply well-established design principles and heuristics, leveraging expert reviews to further reduce uncertainty. Apple might also use A/B testing to compare different design variations by analyzing user interactions, providing valuable data even without direct surveys. Additionally, they review feedback and usage data from previous products to understand trends and preferences. Finally, they make iterative improvements based on all the collected data and expert reviews, resulting in a design with significantly reduced uncertainty, supported by expert knowledge, empirical testing, and historical data analysis.
In this example, Apple reduces epistemic uncertainty through expert knowledge, internal testing, heuristic evaluations, A/B testing, and analysis of historical data rather than direct user surveys. This multi-faceted approach allows them to design well-received products even without traditional user surveys.
Aleatory Uncertainty Example - Managing a Supply Chain During the COVID-19 Pandemic
A company is managing a supply chain for medical supplies during the COVID-19 pandemic. The company needs to ensure a steady supply of masks, gloves, and other essential items.
During the COVID-19 pandemic, a company managing a supply chain for medical supplies faces aleatory uncertainty. Initially, the company plans its operations based on typical demand patterns and lead times. However, the pandemic introduces inherent randomness in demand and supply, which can’t be precisely predicted. Demand for items like masks and gloves fluctuates weekly due to factors like new COVID-19 cases or regulatory changes. Similarly, supplier reliability varies as disruptions cause unpredictable lead times. To manage this uncertainty, the company employs risk management by holding extra inventory, uses flexible contracts with suppliers, and monitors real-time data. This adaptive planning and resilience-building allow the company to better handle random disruptions and maintain a steady supply chain.
In this example, the random fluctuations in demand for medical supplies and variability in supplier lead times during the COVID-19 pandemic are examples of aleatory uncertainty. The company uses system engineering principles such as risk management, flexible contracts, and real-time monitoring to manage this inherent randomness and maintain a resilient supply chain.
By acknowledging and preparing for aleatory uncertainty, the company can better navigate the unpredictable challenges brought about by the pandemic.
Putting It All Together
So, in simple terms:
- Epistemic Uncertainty is about what we don’t know but can learn more about.
- Aleatory Uncertainty is about the natural randomness that we can’t fully control.
INCOSE Handbook: https://portal.incose.org/commerce/store