Overview of the Operations Research Modeling Approach Chapter 2: Hillier and Lieberman Chapter 2: Decision Tools for Agribusiness Dr. Hurley’s AGB 328 Course
Terms to Know Data Mining, Decision Variables, Objective Function, Constraints, Parameters, Linear Programming Model, Overall Measure of Performance, Algorithm, Optimal, Solution, Satisficing, Heuristic Procedures, Suboptimal Solution, Metaheuristics, Postoptimality Analysis, What-if Analysis, Sensitivity Analysis, Sensitive Parameter, Model Validation, Retrospective Test, Decision Support System
Major Phases in Operation Research Studies Define the Problem Gather Relevant Data Develop a Mathematical Model Create or Utilize a Procedure to Generate Solutions
Major Phases in Operation Research Studies Cont. Test and Refine the Model and Procedures as Needed Apply the Model as Needed by Management Assist in Implementing Chosen Solution
Problem Definition Much effort needs to go into understanding the problem at hand. You need to take the vague and convoluted and make it confined and precise. There is a need to understand the appropriate objectives that need to be met. This phase can take considerable time.
Data Gathering Data gathering can take a considerable amount of time. The data might come from primary or secondary sources. ◦ What is the difference between the two? The data may be known with near certainty or could be best guesses (“soft” data).
Data Gathering Cont. Time may be spent conditioning the data. There may be very little data or potentially too much.
Mathematical Modeling A mathematical model is an abstraction of a real world problem which is based on a set of assumptions for the purposes of tractability. It should be noted that when building models, you should start small.
Mathematical Modeling Cont. The main components are: ◦ The Objective Function ◦ The Decision Variables ◦ The Constraints
Create or Utilize a Procedure to Generate Solutions Many algorithms exist for developing solutions for particular mathematical models. ◦ What is an algorithm? Usually these algorithms need computers to find the solution in a reasonable time period.
Testing and Refining the Model Your model should be tested to see if the solutions make sense. ◦ The model may need many levels of refinement to be usable and worthwhile. ◦ It is useful to test a model out with known solutions.
Testing and Refining the Model Most if not all models start out with having issues (bugs). ◦ Bugs should be identified and fixed. ◦ Finding bugs/issues in your model can be challenging. You need to develop a set of skills for identifying issues with your model.