What is O.R.?
What Operational Research Is
Operational research (O.R.) is a way of using analytical methods to help make better decisions. Its methods can be used by almost all organisations, groups and individuals. It uses methods such as logic and mathematical modelling to analyse complex situations, giving decision makers of all types the power to make more effective decisions. Organisations may seek a wide range of operational improvements - for example, greater efficiency, better customer service, higher quality or lower cost. Whatever the organisation’s aim, O.R. can offer the flexibility and adaptability to provide objective help. By using available data it can:
- give a better understanding of what is going on,
- enable a full range of options to be considered, and
- produce precise estimates of outcome and risk.
Furthermore, once better decisions have been made, O.R. people are often central to the implementation of the proposed change.
This site contains sections on various O.R. methods in O.R. Methods, and many examples of O.R. methods in use in O.R. Applications. The next few paragraphs provide some context and linking for the different methods.
We sometimes characterise O.R. as the Science of Better, in other words, using scientific techniques to make things better. To do this, we need a scientific model, and a way of deciding whether things are indeed better. So when we are considering a system, we need three things:
- a model of the system,
- a way of valuing (estimating the value to us of) the outputs of the system, and
- a way of making decisions to improve that value.
These are the 3 components of a typical large O.R. project. But quite often not all three components need to be used. For example if O.R. is applied to a chemical plant, the model structure (the chemical reactions) and the values (market prices of the outputs) may be easily known, so only the decision making rules need to be developed in the study. On the other hand, sometimes in complex management situations, there is considerable value in just clarifying values and objectives, without the need to build quantitative models. This approach is known as “soft O.R.”.
Because O.R. can be used in so many very different situations, there are many different types of model used. However they can be categorised as follows.
First of all, models may be dynamic (showing how the system changes over time), or static, the latter type being mainly used to optimise the allocation of resources to processes or locations. Within both types, there are a range of techniques, depending how far it is possible to quantify the main factors of the model, i.e. to what extent it is possible to represent the factors by simple numbers. The range of countability is represented here:
- all factors representable by numbers (model deterministic),
- variables are simple numerical variables plus an error term (statistical model),
- detailed representation of statistical distributions (stochastic model),
- fuzzy system models, in which variables are represented by sets that represent the possibility or probability of different values,
- non-numerical models, in which logical structures, and causal tendencies are analysed, without assuming numerical values.
Which model type is used will depend on the inherent nature of the system, the availability of good numerical data, and the purposes for which the O.R. analyst is building the model.
In the simplest valuation system, prices can be attached to the outputs, making valuing system output trivial. Often though, this is not possible, for example if components of output deal with quality, or nonmarketable services. In many cases it is still possible to estimate internal prices, and there are a variety of techniques employed in O.R. for quantifying these. This can be done either by deducing implied prices by observing stakeholders’ behaviour, or more directly, employing a variety of survey methods to ascertain stakeholders’ stated preferences. In all these cases, the existence of prices for each of the various outputs means it is possible to sum together the various outputs into a single figure of utility for the aggregate output.
But this is not always possible. Sometimes the different outputs are incommensurable, so O.R. has a number of methods that deal with multiple criteria, often multiple objectives. Techniques also exist for putting these criteria into structures, such as hierarchies. A further complication occurs when there are multiple stakeholders, so methods have been developed for clarifying, and hopefully making consistent, their different viewpoints.
Given a model of a system, and a system of valuing output, there is a ideal theoretical way of making a decision to make the system better, which is to optimise the value of output, or, if only objectives have been defined, finding a solution which meets all objectives. The O.R. solution can then be directly implemented. In many management situations however, this is not practicable. An organisation may have a limited set of acceptable policies, where a policy is a set of objectives, together with a preferred way of achieving them. Such a management situation cannot be dealt with just by giving an optimal solution. The management have to be involved in the decision process. There are many ways of dealing with this, ranging from real time simulation or gaming, with the manager involved in the modelling process, to the O.R. analyst running a range of scenarios, so that the choice can be left to the manager.
O.R. Methods and O.R. Applications
There are a very large number of O.R. methods covering modelling, valuing and decision making. We intend gradually to expand the section on O.R. Methods on this site to give examples of many of these.
The subject matters of O.R. are also manifold. Different factors of production can be concentrated on, such as :-
- capital, and equipment
- other input resources, for example energy
The section on O.R. Applications contains many articles organised by topics, such as these factors, or by industrial sector (such as health and defence). It is aimed once again, eventually to provide a comprehensive set of articles for the use of O.R. practitioners.
For more About O.R., take a look at our other websites:-
This site’s title, Learn About O.R.
, says it all – it contains help for anyone who wants to know what O.R. is, how it helps organisations, and how to get into the profession. It aims to encourage youngsters to study maths, and provides teachers with resources to help teach O.R.-related maths topics.
The Science of Better campaign is designed for senior executives in the business community, to give them a better understanding of O.R. It explains what it is, how it’s so powerful and how it could help their businesses to do better.