Online demonstrations
 An important part of the Decision Analysis Initiative (DAI) will be the setting up and running of online demonstrations of decision analysis applications. These will be supported by additional information in the form of help systems, didactic material, audio-visuals, briefs and other relevant documentation. The initial demonstrations are being assembled so as to present a graded sequence covering successful applications illustrating aspects of functionality as well as integrated systems. Demonstrations will include core decision analysis models used to explain the objective and methods applied according to the knowledge domain.
Functions are not enough
The recent upsets in the world of finance resulted in a recognition that decision analysis and the assessment of risk are not simply matter of applying a sophisticated mathematical function to "similar" circumstances. The effective application of statistical routines and mathematical functions is completely dependent upon the precise logical relationships and probabilities of the critical events affecting outcomes of decisions. This objective approach can only be sustained if the reality of constant change is recognised. In decision analysis, an "average" reference value, based upon previous events, can be of little relevance to a current decision affecting future outcomes. Such changing surrounding circumstances causes each decision analysis to be unique, each is one of a kind, perhaps similar to - but never identical with - previous situations. This has always been recognised since the emergence of decision analysis as a new discipline in the 1960s. Early work by Matheson & Howard1 explained that decision analysis seeks to apply logical, mathematical, and scientific procedures to the decision problems that are characterised by uniqueness. The uniqueness is accommodated in what is known as the decisiona analysis model, a logical framework holding the relevant components, such as data-intensive functions as well as relating these to the data inputs and outputs.
Besides this need for adaptability in decision analysis models, the other typical characteristics include the importance of decisions which often place a significant portion of an organization's resources in question. Uncertainty predominates with many of the key factors that must be taken into account being imperfectly known. Decisions can have long run implications, an enterprise can be forced to live with the results of the situation for many years, perhaps even beyond the lifetimes of the individuals involved. Decisions usually need to satisfy complex preferences requiring a transparent incorporation of the decision-maker's preferences about time and risk. Decision analysis needs to provide a practical computational implementation of a decision analysis model by configuring the relationships between analytical components to reflect the decision-making environment. This helps achieve a more objective and flexible mathematical modelling of the decision and more realistic quantitative evaluation of the various courses of action.
Demonstration technology & techniques
Taking into account the need for a decision analysis model building capability, as well as having access to powerful analytical decision analysis modules, the DAI needs to build demonstrations making use of tools which combine the following properties:
- an advanced decision analysis scripting language with a sytnax compliant with international standards such as ISO or ECMA
- server end processing capability for speed
- compatible software tools including:
- script coding system or author capable of building decision analysis models and placement of relevant functions
- a script with an appropriate-sized list of relevant processing functions or primitives
- an ability to create reusable applications primitive libraries with the script
- visually intuitive modelling system providing logic "lay-outs", flow structure and access to functional detail when necessary
- a language compatible with and supporting inclusion of scripts complying with international quality standards
- generated code is compatible with server protocols and immediately executable
- cross compatibility of server output with different client browsers
- cross compatibility of server output with different client end operating systems
|
Adoption of the ECMAScript core standard
 |
Taking into account these requirements all demonstrations on the DAI site will be based on server end JavaScript operation using the implementation of the ECMAScript core standard (established standard by ECMA and the ISO). To provide this functionality use will be made of DScript® a scripting language developed by Vanguard Software Corporation. DScript® is a server end implementation of ECMAScript with a ten year operational track record of successful applications. DScript® provides a transparent means of using standard ECMAScript for server end operations. ECMAScript can be combined with DScript in the authoring platform (Studio) and this runs efficiently, with no modifications, as a server end implementation. DScript® has around 1,000 "primitives"of which 50% are documented and 50% not. Some 500 of these are powerful integrated functions providing most of the effective operations research, logic, statistical and mathematical functionalities used in decision analysis. Use will also be made of the Seel-Telesis Script Library made up of some 2,000 functions, modules and primitives covering vertical and horizontal applications. The Seel-Telesis Script Library is all written in DScript and ECMAScript. For client end operating system control, via the online client platform, use will be made of the Seel-Telesis OS Script Library. This has around 3,500 modules which manipulate the Windows® operating system as well as client end Windows® applications in silent, explicit and automatic modes both in offline and online status.
Contributing to advancing applications & standards developmentUnder the DAI will be running a range of demonstrations which will include benchtests of programming efficiency. These two aspects are inter-dependent in that as applications fan out into different areas the constraints and avantages of scripts used become apparent. We will provide regular feedback to the George Boole Foundation's own standardization work so as to identify the most important development needs as well as the most promising routes.
1 Matheson J. E. & Johnson R. A., "An Introduction to Decision Analysis", Stanford Research Institute, 1968, cited in McNeill H. W., "The State-of-the-Art & Future of Decision Analysis", Boolean Society, Brief 001, January, 2010.
This site is sponsored by SEEL Systems Engineering Economics Labs http://www.seel-telesis.com
|