Operations Analysis (OA)
Operations Analysis describes a set of analytical tools and methods that assist in determining the optimal ---that is---the most efficient and effective allocation of scarce resources.
A pamphlet distributed by a leading Operations Research (OR) organization, INFORMS, titled Seat-of-the-Pants-Less, describes OR/OA as, “…the discipline of applying advanced analytical methods to help make better decisions.”
Practitioners of OA use three primary analytical tools: data analysis, mathematical programming, and modeling and simulation.
Data Analysis
Data Analysis begins with a dataset and attempts to determine which input factors are having an effect on the outcome. We use this technique for making predictions, identifying trends, and finding those variables that contribute to the desired outcome. Data analysis is easy to perform and can be done quickly given a historical dataset.
For example, we may have 1,000 instances of bearing failure and for each instance we have identified several related factors such as: the number of hours of operation; the heat of the bearing at the time of failure; the rpms of the bearing at the time of failure; and the amount of degradation the bearing had sustained as of the last measurement. From this dataset we can determine, with a specified level of confidence, which of those factors is ‘important’ in forecasting future bearing failures.
Mathematical Programming
Math Programming, or Optimization, seeks to optimize (maximize or minimize) a specific objective given certain resources and constraints. Objectives are usually stated in terms such as 'Maximize readiness' or 'Minimize Cost'. Resources are things such as the amount of labor hours available or the amount of cubic lift we have. Constraints limit how those resources must be used. We may specify things like, ‘within budget’ or ‘by the Required Delivery Date’. The mathematical programming process then translates those objectives, resources, and constraints into an algebraic expression that can then be calculated.
Modeling and Simulation
Modeling and Simulation (M&S) seeks to recreate a process with a set of computer instructions (a model) and then repeatedly vary designated inputs (simulation) in order to explore how those inputs affect areas of interest. M&S allows us to quickly and cheaply explore many variations and outcomes of a process to determine the ‘best’ combination of inputs. The M&S process will produce a dataset that we can then use data analysis techniques on. We use this technique for network analysis, resource allocation, and sensitivity analysis. Optimization requires input from Subject Matter Experts. M&S is used to explore and refine processes prior to investing capital in full scale production. The level of detail in models can vary widely and depends on input from Subject Matter Experts.