How to Study for an Exam in Operations Research
J.W. Chinneck. April 19, 2021.
Exams in Operations Research or Optimization normally require you to formulate and solve word problems in exactly the same manner as on the assignments. This means that for full marks you must include a complete formulation, a solution, and a concluding statement (unless otherwise stated).
Many students find that they are pressed for time on such exams. Time management during the exam is vital.
How to Maximize your Marks on Any Exam
- This should be obvious: study the material. For a problem-solving course this means practicing solving problems. Passively reading over your notes is not an effective way to prepare.
- At the exam, take a couple of minutes right at the beginning to skim over the question paper quickly, maybe putting a tick mark beside the questions that you are sure you know how to answer.
- Answer the questions with the tick marks first. This way you'll get all the "easy" marks first and if you run out of time, it will be on the questions you're not sure how to answer.
- Calculate a time budget for each question and stick to it. For example, if the final exam is three hours long and has 6 questions, each of equal value, then you should budget 30 minutes per question. Getting hung up on one question can be costly: move on to another question when the time budget for the current question is exhausted. Come back to unfinished questions later if you have time.
How to Study
- Make sure you understand where and why you lost marks on the assignments. The examination is exactly like a collection of assignment questions, so you should be able to do those assignment questions well. Of course, the limited number of assignment questions means that not every solution technique was exercised in the assignments, so you should also look at other sources of worked examples.
- A course of this type covers many different techniques and the biggest cause of lost marks on a final exam is applying the wrong method in an attempt to solve a problem. Up until now it has been easy to know which solution method to apply because the questions in the assignments followed right after the material was presented in class. On the final exam it's different: you don't have any external clues about which method to apply to each question, so you need to practice identifying which method to use on any given problem. Here's a good way to do that:
- Organize yourself into groups of 3-4 people.
- Each person in the group should find a selection of problems, e.g. from an Operations Research textbook, or from online sources, for which you know the right solution technique to use (e.g. because you also have the solution, or because you know which chapter in the book it was taken from).
- Each group member presents their questions to the rest of group, who don't know what type of question it is. This provides everyone with practice at identifying the type of solution method to apply to previously unseen questions.
- Pay particular attention to differentiating branch and bound questions from dynamic programming questions. These are often confused, resulting in the loss of many marks. How can you tell which is which? Some clues:
- It's very hard to find a bounding function that has any look-ahead value for a dynamic programming problem.
- We looked at two main types of dynamic programming problems: equipment replacement problems, and knapsack-type problems. If the problem matches either of these forms, then it is likely a dynamic programming problem.