Linkedin

Operations Research

Operations research (OR) is less familiar to most clients than our other offerings.  In brief, it refers to analysis of the operations of an organisation (in its origin, these were military operations in WWII).   The OR analyst will determine the key elements and interrelationships of these operations, form a model that describes the system, (often a mathematical model), and then use the model to improve or even optimise operations.

Specific mathematical techniques have been developed to address types of problems.  Some selected types of problems and the corresponding operations research technique include:

  • Effective allocation of limited resources to competing demands:
  • various forms of mathematical programming, such as linear programming, integer programing, and dynamic programming
  • Provision of adequate service points to meet fluctuating demand:  queuing theory
  • Planning of geographical networks of service delivery (within a building or across the country): network analysis
  • Developing strategies to cope with competition: game theory
  • Reducing investment in inventory while limiting the chances of running out of supplies: inventory theory
  • Design of organisations that balance efficiency with robustness:  reliability analysis.

On some occasions, straightforward algebra and calculus is sufficient to solve the problem.  On other occasions, the problem is not amenable at all to mathematical analysis, but can be addressed by simulation techniques.

Previous clients (of the director)

Department of Defence - Navy, Army, Air Force and the Department, with studies involving logistics, operations, strategy and personnel.

Delivering an operations research study

The operations research approach requires a significant degree of trust between the sponsor and the analyst. The mathematical model is necessarily something of a 'black box' so the analyst needs to consult closely on the assumptions that go into the model.  One of the main advantages of an OR study is that it can produce unexpected results that lead to considerable improvements to current practice.  But to gain acceptance of such non-obvious recommendations, the analyst must be able to explain in clear terms to all stakeholders why the model produces the results that it does.