Replies to This Discussion

Permalink Reply by arup guha on April 21, 2009 at 12:47am

I thought a good thumbrule as to what is stats and what is OR is that statisticians know about distributions more than anybody else. Algorithms etc would be more from the data mining world and not from OR. and math modeling is for the applied mathematicians. i thought the OR specialists deal more with optimization kind of problems. and statistics is most certainly not a part of OR, no matter what the OR professionals like to think. Tomorrow they would claim that pure maths is in some way a prt of OR so pure mathematicians are actually OR professionals in some convoluted way.

Permalink Reply by Yi-Chun Tsai on April 21, 2009 at 7:54am

I have degrees on both statistics and operations research but now am working as a marketing statistician. From work point of view, there are not many opportunities out there that require background in both fields, even though I think having background in both fields would certainly make you more marketable in terms of job searching.

For example, when my company wants to do inventory management, they use software that first forecast the demand for each product before they plan on safety stock level. In this case, if you can do statistical forecasting as well as optimization, then you certainly can do the job. The thing is that in my company, they have two positions for this task.

Permalink Reply by Deborah Deng on April 28, 2009 at 2:56am

In my opinion, the two fields are quite different because their goals are not the same. Statistics are used to describe problem and predict future etc.; the goal is to extract information. Yet, we apply methods in Operations Research with the information to decide. However, the practitioners in these two disciplines often need to borrow the tool kit from each other. I guess this is how they become inseparable. In my experience as an OR analyst, the first thing I always do is to collect data before I can start building any model. From there, I have to use all my statistics knowledge to obtain the most accurate parameters to use in my model so the result will be realistic. That is probably why there are heavy Statistics courses in OR program. I believe there is the same thing for Statistics; when encountering a task to decide, OR skills apply.

I think OR and Statistics are two fields that one can choose one and dig very deep into; but one can also have a lot fun applying both of them to solve real-life problems. We are in the world of creating more and more cross-discipline theories to have lots of idea and lots of fun. :D

Permalink Reply by Vishrut Talekar on August 31, 2009 at 12:33pm

I think you can say Operations Research is field of Applied Math. The deeper you go more MATH you need.
Also there are areas like Markov Chains, Stochastic which both OR and Statisticians follow.
Overall I feel its all Applied MATH with different categories. And It really doesn't matter what name you give.

Permalink Reply by Kevin Chen on September 1, 2009 at 8:22am

Agree. It's all math, under different hats.

Permalink Reply by Sarat Patro on September 20, 2009 at 10:29pm

Having trying my hands on both of these fields this is what I feel : we need both stats and OR to solve a good many no. of business problems. However Stats is more on the stoachastic side and not deterministic whereas OR is more exact . There will be only one OR solution to most of the OR problems if we agree on certain key points regrading how to execute. Stats on the other hand is more of multiple solutions to the same problem. Tell that there is only one type of model possible to any stats guy and well we have taken out the fun out of modeling..

Permalink Reply by Chandrasekhara S. Ganti on October 7, 2009 at 8:19pm

Statistics is more like after the fact data analysis. Of course, you can also design a study, collect the data and perform analysis. Optimization in Statistics includes the theory and procedures like MLE. However, Forecasting and Predictive Modeling is still for the future, but the major difference in Stats and Operations Research is :

Operations Research for most part is concerned with ex: Maximizing (Profit) or minimizing the Cost an objective that is subject to constraints on available resources.

However, Statistics is concerned with descriptions, predictions, where most part there are not any constraints nor an optimal value is sought even in the prescriptive models, Optimal Forecasts , Optimal Classification in Discriminant Analyses-- you may have heard which means optimization calculus is always involved behind the scenes.

Where the two meet: in Queuing Theory, Inventory Theory, Reliability theory: Statistical Distributions of inputs and modeling; Simulation where Random Samples of Distributions are used to Simulate the whole process.

A note on F and other values in SAS is a classic example: in SAS / OR modules and in there exactly you need to find if a Supply Chain, Demand, Inventory inputs or output distributions and modeling the whole processes.