What is an actuarial analyst

what is an actuarial analyst

PROFILE OF AN OPERATIONS RESEARCH ANALYST

The OR analyst applies scientific method to problems concerning the management of systems of people, machines, materials and money in industry, business government and defence. He or she conducts logical analysis of management problems in collaboration with management with a view to understanding the system behind that problem, so that the system may be made to work in a manner which eliminates the problem.

The analyst will strive to define the objectives and constraints of the system in quantitative terms and will generally develop a mathematical model which includes chance and risk factors. The analyst often operates as an agent of change within the system and must therefore develop interpersonal skills and be able to work with management as part of an interdisciplinary team.

Standard OR techniques such as Linear Programming, Critical Path Analysis, Simulation, Statistical Decision Theory, Queueing Theory and Inventory Control Theory may be used. Specialist work may be in the production, marketing or financial planning areas.

OR work can be with any one of a large number of industries, including oil, chemicals, steel, manufacturing, the railways, electricity generation, agriculture, the airlines, banking, insurance and many other fields including defence, medical and welfare services.

DOES OR HOLD A FUTURE FOR YOU?

If you do decide to follow a career in Operations Research your future is very much in your own hands. Initiative, flair and creativity are just as necessary as integrity, ability and hard work. It is a profession which is rapidly developing.

Most OR projects are handled by small project teams and in your early years you might be a junior member of such a team. However, it would be unusual if during this period, you did not also have some projects of your own. Many young graduates become project leaders within two years. It is really up to you. Quite often you may be working on several projects at the same time, in a different project team in each case.

The ideal OR analyst needs to be able to identify the right problem to solve, collect the relevant data, construct a model of the system often using a computer, and seek a solution that people in the real world can live with. Finally they may need to "sell" their solution to management.

Generally, you would be able to move around as many different organisations employ OR personnel. It is not unusual for OR graduates to gain experience in other countries and they are frequently required to travel within large firms or government departments. This is certainly true if you join a consulting firm.

WHAT IS OPERATIONS RESEARCH

Operations Research is concerned with analysing complex problems and helping decision makers work out the best means of achieving some objective or objectives.

Thus Operations Research requires an enquiring mind capable of asking probing questions to help clarify the decision maker’s objective. Within a business organisation the primary objective will often be either:

  • Maximising profits, or
  • Minimising costs

This may seem simple, but to come up with a good solution to a problem it it necessary to identify ail of the factors which contribute to costs or profit and that is not so easy. In many situations, more than one objective is set, and these will typically be in conflict.

Further, Operations Research is not only used in business: in other applications the real challenge of the work is often concerned with quantifying objectives which cannot be expressed in financial terms.

Any proposed solution to a problem must be practicable and take account of any constraints which limit the range of possible actions. Thus a systematic approach is needed to identify all aspects of a problem and analytical ability to work out how the different parts of the problem situation are related to each other. Finally, the analysis must take account of risk and uncertainty.

HOW OPERATIONS RESEARCH IS CARRIED OUT

Most OR starts with a problem, which does not necessarily mean something has gone wrong or is about to go wrong. It could just be that a decision has to be made or that the person or group responsible for some activity believes it could be carried out better: that there is scope for improvement. From that start, there are a number of more or less standard steps to the conduct of the investigation.

Problem Definition

Obtain as complete a picture as possible of the situation being investigated and identify how all the people involved see the problem or decision. Then identify alternative courses of action which are available to solve the problem or improve the operation. Not infrequently, the problem definition will lead to a quite different view of a problem or of the scope of a decision from that originally reported.

Specification of Objectives

Develop a clear statement of the goal or purpose of the organisation or activity being studies. This is a crucial step, and never easy. Do not assume that there will always be one person who will immediately give you a clear, unambiguous statement of a single objective. Part of the problem will usually be that different people are responsible for different aspects of an activity, and that each sees the objectives in a different light. So, the OR analyst will need to talk to all the people concerned and collect all their views on the goals and objectives of the activity, identifying any conflicts between different objectives and then trying to find some means of resolving these conflicts and constructing an overall measure of performance. Sometimes, on reflection, the overall objective can be specified as the best compromise; in other cases it is not so easy to see how to balance one objective against another.

Identification of Constraints

Find the features of a problem or activity which limit the range of possible options, or alternative courses of action. We have already mentioned budgetary constraints - limitations on the amount of money available to spend on an activity. With manufacturing problems, there will usually be constraints associated with plant capacity and quality requirements. Sometimes, when there is more than one objective, some of the objectives can be converted into constraints by specifying minimum acceptable levels of performance. Sometimes, the constraints are found not to be as absolute as was believed, then the OR specialist can contribute by demonstrating how to relieve such limitations.

Data Collection

Once all of the characteristics of the problem have been identified, the next difficult task is to accurately quantify all items. The search for complete, correct data is a difficult, but an essential task if a valid result is to follow.

Model Validation

Before a model can be used as a decision tool it must be validated. This is usually carried out by demonstrating that the model can accurately represent the known prior behaviour of the system.

Evaluating The Alternatives

Find how the alternatives measure up against the objectives. If there is a single easily measurable objective such as cost or profit and a finite list of alternatives this can be straight-forward. However, objectives like human well-being are not so easy to quantify, so such cases can require a lot of careful thought and judgement.

Selecting The Best Solution

Choosing the alternative which measures up best against the objectives. If the analysis comes up with a single overall objective which is acceptable to all people concerned the choice many follow simply from the previous step. However, if there are several objectives which are difficult to combine into a single overall objective, then the analyst can produce a short list of good alternatives which look good against one of the objectives and not too bad against the others. Thus often the solution will involve compromise amongst conflicting or competing objectives.

Implementation

Getting the people responsible for the problem or activity to act on the recommended solution, or - if there is not a single recommended solution - to use the results of your analysis to make a well informed decision. Indeed, being able to favourably influence the framework of decisions is often of greater long term value than getting a particular recommendation accepted.

Monitoring

While it is always nice to find that a proposed "solution" to a problem works in practice, it does not always work out that way. OR work cannot truly be said to be complete until it has been demonstrated that the implementation achieved a satisfactory improvement in performance.

OPERATIONS RESEARCH AND MATHEMATICS

Much Operations Research makes use of mathematical concepts, but from what you have already read in this booklet it will be clear that Operations Research is more than just mathematics. Sometimes the mathematics is just addition, subtraction and multiplication, while other projects use more advanced mathematics and a computer is required for the computations. However, even when a computer is used, the basic mathematical concepts are often quite straight-forward, but would take far too long if done by hand.

Similar mathematical methods can be used for many problems with a common mathematical structure and this has led to the development of some standard OR TECHNIQUES.

Examples

are:-

Network Analysis

The inter-related activities in a complex situation (say, a construction project) are represented by arrows on a diagram. The diagram shows the logical relationships between activities, and, once it has been drawn it is fairly simple to determine the "critical path': that sequence of activities which determines the total duration of the project.

Linear Programming

One of the earliest developed and most widely used OR techniques. One of the main applications is in planning manufacturing activities. The elements of a linear programming problem are:

  • Variables:
    • Quantity of each product to produce
    • Quantity each raw material to buy
  • constraints:
    • Market demand for the products, plant
    • Capacity and product quality
  • Objective function:
    • A mathematical formula which expresses the objective say profit, as a function of all the variables.

The technique is called linear programming because it can only be applied if the objective function is a linear function and the constraints are linear equations or inequalities.

Stock Control Theory

Another standard OR application, which includes a mathematical method for finding the best compromise between the periodic cost of ordering or producing an item and the cost of carrying stock. The natural preference of a production manager may be to choose long production runs so as to spread set-up costs over a larger number of units, while an accountant may want to have short production runs so as to minimise the costs of holding stock. Stock control theory provides a method of selecting a quantity which minimises the sum of these two costs, and the formula can be developed using fairly simple calculus.

Statistical Analysis

Is not so much an OR technique as an aspect of mathematics which forms an essential part of the background training of an O. R. analyst. There is a lot of uncertainty associated with most real life problems, so the OR analyst needs some knowledge of statistics to be able to handle that uncertainty in a rational manner. However, what is often needed is not so much a knowledge of specific statistical techniques as a general ability to think in statistical terms.

The mathematical basis of all of the above techniques can be understood by someone with a good grounding in High School mathematics. However, real problems tend to be bigger, messier and more complex than text book problems and therefore an operations research analyst needs to:

    • exercise judgement and creativity in adapting standard techniques to accommodate all aspects of a real problem.
    • use a computer to handle the sheer volume of often repetitive calculation.
    • have qualities of perseverance and patience to see a big project, through to successful completion.

The available techniques, of which the above are a few examples, is constantly growing. Thus this is an active, evolving field of study offering opportunities for the enterprising person.

THE HUMAN SIDE OF OPERATIONS RESEARCH

Having explained something of the connection between Operations Research and Mathematics, it is important to emphasise the human side of Operations Research. As mentioned in the general description a very important part of OR is concerned with talking to people about a problem; getting them to describe the objectives and constraints. This requires a lot of personal skills:-

  • ESTABLISHING THE RIGHT SORT OF FRIENDLY BUT BUSINESS LIKE ATMOSPHERE IN WHICH TO CONDUCT YOUR DISCUSSIONS.
  • COMMUNICATING CLEARLY THE PURPOSE OF YOUR INVESTIGATION.
  • LISTENING CAREFULLY AND ATTENTIVELY, SO THAT YOU PICK UP THE CORRECT EMPHASIS OF WHAT THE OTHER PERSON IS SAYING AND THINGS WHICH ARE IMPLIED BUT NOT SAID DIRECTLY.
  • JUDGEMENT IN BEING ABLE TO ESTABLISH THE TRUTH WHEN YOU GET CONFLICTING ACCOUNTS OF THE SAME "FACTS" FROM DIFFERENT PEOPLE, THUS AVOIDING JUMPING TO PREMATURE OR ONE-SIDED CONCLUSIONS.
  • WILLINGNESS TO LEARN THE TECHNICALITIES OF FIELDS ENTIRELY NEW TO YOU, SO THAT YOU CAN UNDERSTAND WHAT IS REALLY HAPPENING.
  • ABILITY TO ACCEPT AND DEAL WITH INPUTS THAT ARE VAGUE, UNCERTAIN OR UNQUANTIFIABLE.
  • SENSITIVITY TO THE FEELINGS OF PEOPLE WHO MAY BE RESISTANT TO OR FEEL THREATENED BY YOUR INVESTIGATION OR RECOMMENDATIONS.
  • HELPING TO BUILD CONSENSUS AMONG A DIVERSE GROUP OF PEOPLE AS THE BASIS FOR EFFECTIVEIMPLEMENTATION.

OR EXAMPLES

Integer Programming At School

Every school is run on a master time-table. The pupils see only parts that affect them - the times, subjects and teachers and rooms. The teachers see a different time-table, with classes, subjects and rooms. Consider a school of say 1 000 pupils in about 40 classes ranging from year 7 to year 12, with about 80 staff and nearly sixty rooms counting special laboratories etc. Then think how you would allocate teachers and classes to rooms to maintain adequate supervision of all classes, to avoid clashes over rooms, to observe the working conditions for teaching staff, but primarily to provide a properly balanced teaching programme.

If this objective of a "balanced teaching programme" can be specified in the appropriate quantitative terms the problem can be represented as a linear programme with all of the requirements of avoiding clashes over rooms and so on as constraints. However, it is a special type of linear programme, an Integer Programme with variables required to be integers: a classroom at a given time has to be allocated to one class or another; it can not be allocated half to one class and half to another.

Linear Programming For Refinery Planning

Shell Australia operates two refineries, one at Geelong and the other at Clyde in the Sydney metropolitan area. The major inputs to these refineries are (-rude oil from Bass Strait and crude oil imported from Indonesia and the Middle East. The crude oil goes through a number of processes such as fractional distillation, which separates out the light components from the thicker heavier oils; cracking, which breaks down the heavy constituents of the crude oil into lighter components'. reforming which changes the chemical structure of other components in order to meet product specifications; and desulphurising which removes sulphur from the product in order to meet environmental and quality requirements. That list is far from complete, so it will be evident that an oil refinery is a very complex affair which can be operated with considerable variation mix of crude oil inputs. However, the job of the refinery economist is to work out the best way to run the refinery to decide how much of each crude to run through the distiller at each refinery: how much of the heavier fractions to run through the cracker, and so on, in order to find the most profitable way of meeting the expected market demand for:

  • petrol
  • diesel
  • kerosine (for jet aircraft)
  • fuel oil (for industry)
  • bitumen
  • lubricating oil and other products.

Up to about 25 years ago all of this had to be planned on the basis of experience and trial and error calculations, but now Shell (and other major oil companies) use linear programming to solve this problem. This is possible because profit can be expressed as a linear objective function and linear equations or inequalities can be used to represent all of the constraints concerning:-

  • the available quantities of different crude oils
  • quality requirements for each product
  • market demand for each product
  • Two of the biggest advantages of using LP are:-
  • When circumstances change significantly it is possible to produce a revised refinery plan much faster than previously.
  • The linear programme can be used to investigate the impact of anticipated future changes and to check that a proposed expansion of one part of a refinery gives a suitable overall improvement.

Production Scheduling In A Shoe Factory

When shoes are made in a factory the manufacture of each type and size of shoe requires the use of a number of machines. The different types of shoes are made in batches of several hundred and each type of machine has to be adjusted or "set up" for each batch of shoes. At any given time there will be a number of types and sizes which need to be made in the immediate future and a decision has to be made as to the sequence in which the different batches pass from machine to machine. Any sequence chosen is likely to mean that for some of the available time some of the machines and operators will be idle. Hence, it is desirable to choose a sequence which minimises this idle time.

The problem was tackled for a Victorian shoe manufacturer by two students as a project which formed part of the final year OR course at an Institute of Technology.

The problem they studied is illustrated by the following example:-

Consider eight batches of shoes of types A, B, C, D, E, F, G and H which must pass through four sections during manufacturing with the following times required (in minutes) for 600 shoes of each type in each section,

Source: www.asor.ms.unimelb.edu.au

Category: Insurance

Similar articles: