• home
• our services
• our experience
• our customers
• geek
• about us
• contact us
• log in

Projects with which we have been significantly involved with include:

Defence policy (optimisation)

Policy and policy related discussions are often performed without the benefit of quantifying the economic cost and benefit of numerical optimisation.

Policy is often expressed in very loose terms; this makes it very difficult to ascertain what is within policy constraints and what is not. Policy might imply a outcome, but not the extent of that outcome. By example, a policy related to Defence might be to do Search and Rescue, but it will not for instance say that it should be able to handle all the Search and Rescue emergencies. The following problem was designed to produce the maximum "Value for Money" from the Defence asset portfolio.

Policy is set by Government, often as part of a "White Paper"; in general terms it lays out the Government's objectives.

In the following, we have chosen not to use the term "capability" as it is often loosely used to mean an effect such as achieving a specific Task, or the Means to achieve that effect. For the sake of this discussion, Means is used to achieve an effect; and Task is used as the effect.

Modelling Approach

A four state modelling approach creating distinct scenarios was undertaken as follows:

1. A set of all possible Means was produced. Each Means was defined and costed. The set was closed, the addition of a new Means involved the definition and creation of a new scenario;
2. The set of all Policy decisions (Tasks) was created. The set was closed ie there were no Tasks outside this set. New Tasks when generated became components of new scenarios. The Tasks included options that might have been valid Tasks that were not currently government policy. Each Task set was defined with sufficient rigour so that associated decisions of Means could be tested. Each Task in a scenario was valued on a consistent basis. The value tentatively used has been the loss that would incur if no action was taken. The value was then weighted by its frequency of occurrence, and the probability that government would engage. Importantly, the Defence Force exists to achieve outcomes valued by the Minister. The value paradigm was set up to catch a Ministerial view on outcomes. It did not capture military centric view.
3. Rules were established relating the Means required to achieve specific Tasks. Importantly some of the more expensive capabilities eg air transport were common to many Tasks. These rules are as simple as to achieve Task 1 it is necessary to have 2 units of Means 'A', or 2 units of Means 'D' and 1 unit of Means 'E'.

At this point, it is possible to set a fixed budget of aggregate CAPEX and OPEX required to operate the Means, then explore the combinations of Means not exceeding the fixed budget(s) that created the most aggregate value in terms of achievement of Tasks.

The optimum portfolio was able to be determined. The optimum portfolio provided a set of Means costing no more than a upper cost limit while generating the most value of Tasks completed. This is a Pareto optimal solution. No other set of Means chosen with the same upper cost limit will produce greater value of Tasks completed.

Analysis Delivery

A website was constructed allowing multiple parties to explore the value generated through capability decisions. The optimisation exploration was achieved using sliders on a web page. The sliders were moved and the impact on policy was immediately obvious in terms of the cost and the benefit. Sliders allowed a manual exploration of cost v. benefit trade-offs. Once the relationship between Means and Tasks were setup, then manual and computer based optimisation was possible. Computer based optimisation considered any constraints specifying upper or lower limits on capability, and any required outcomes. It then produced a recommended Means set to maximise the value of Tasks able to to completed without exceeding the various cost constraints.

Benefits

This was a paradigm shift. It produced a mathematically rigorous transparent optimisation of the defence capability. The benefits were fully traceable back to Ministerial requirements. Budgets were not exceeded. One of the most major benefits was that the consequence of capability scenarios could be investigated in minutes, as opposed to days or weeks using manual techniques.

Informally, we noticed that a computerised optimisation would produce often 50% more value for the same budget that the human attempts at optimisation. This has produced the same value of Tasks performed for half the cost, or alternatively producing twice the value from the same input budget.

At no time would it be considered that strategic planning should be performed by computers, however the prototyped system provided a valuable means of communicating a range of complex decisions to various parties.

The top three skills required and used for this project were:

• Workshopping Requirements
• Economics
• Modelling