Chapter 1 - Introduction Flashcards
WHat is operations research?
Operations research is basically research on how operations is to be done. This means, how activities and resources should be executed.
Operations research is about coordinating and conducting operations within an organization.
Elaborate on the “research” in operations research, and point out the important parts
The research part suggests that the OR field use an approach that resemble the scientific method. (Recall scientific method is about testing a hypotheses through multiple experiments).
In fact, there is a lot of equals between OR and scientific method.
Broadly speaking, we follow these lines:
1) Carefully identify the problem and formulate the problem.
2) Construct a scientific (typically mathematical) model that attempts to abstract the essence of the real problem. by this, we hypothesize that the model is accurate enough to produce solutions that are valid for both the model and the real problem.
3) We apply experiments to the model, testing its capabilities.
4) We modify the model is the experiments say so
5) Eventually (hopefully) we verify something.
6) Provide clear and instructive results that can be used in the organization.
What is the difference between descriptive, predictive and prescriptive analytics?
Descriptive analytics analyze what has happened in the past or in the present using data.
Predictive analytics use models to try to predict what will happen in the future.
Prescriptive analytics use models to create/advise managerial decision making.
The book focus mainly on the mathematical methods and algorithms etc. Due to this, what is a dangerous pitfall?
OR studies includes a lot more than just the mathematical work.
We need to define the problem, gather data, analyze the data (descriptive analytics), use predictive analytics, formulate math model, derive solutions, test the model, prepare to apply the model, implement the model.
So, the math-part is only a small part of the problem.
Which of the 3 categories is typically the one that OR focus on?
The prescriptive analytics. Telling/advising managers what to do.
OR also has a little (not much) on predictive analytics.
However, OR is barely on the descriptive analytics train. It is common for organizations to hire individual data scientists that specialize in gathering and analyzing data, which takes care of the descriptive analytics. These also takes care of the predictive analytics. Thus, OR concerns with the prescriptive analytics.
What are the 9 phases of a full-fledged study conducted by OR and analytics professionals?
1) Define the problem of interest
2) Gather and organize data
3) Use descriptive models to analyze the data
4) Use predictive models to analyze big data
5) Formulate a mathematical model to begin applying prescriptive analytics
6) Learn how to derive solutions from the model
7) Test the model
8) Prepare to apply the model
9) Implement the model
Consider some objective function. In what way do we incorporate the data we have gathered into this objective function?
Objective function could look like this:
P = ax1 + bx2 + … + wx22
The constants would represent values coming from the data. We call them “parameters” of the model.
Is a solution to an optimization problem optimal?
It will be optimal in that model. However, since models are abstractions of reality, we cannot guarantee optimality in all cases. Therefore, we must test our models.
Herbert Simon said…?
He pointed out that “satisficing” is much more important than optimizing in actual practice.
What is the benefit of satisficing?
Let us say we are in an org that has a lot of decision variables becausei t has a lot of different operations.
If we want to optimize, we would be looking for the overall shite that optimize between conflicts. Instead, we can establish satisfactory goals for the different operations, and then solve using these boundaries instead.
What is the roots of OR?
Formally it started during WW2. However, it was first used by organizations shortly after as an attempt to bring the scientific approach to managing organizations.
What are the two fundamental reasons for the boom in OR?
1) Much progress on the OR methods, like LP etc
2) PC revolution, making processing power very available.
