Community Research and Development Information Service - CORDIS

Improving means for problem solving

Optimal decision-making is about empowering people to take the best possible decisions. Currently, few tools are available for assisting decision-makers with this highly complicated task and those available have restricted areas of application, not to mention that they require highly specialised backgrounds in decision theory. The Optimal Decision Making application package has been developed to give users more flexibility to maximise their own efficiency in testing hypotheses for selecting and implementing the optimum decision.
Improving means for problem solving
Decision-making is the actual process of selectively judging when presented with several complicated alternatives that consist of several variables. Most cases of decision-making problems are scientifically grounded on the application of stochastic mathematics and statistics on experimental results, which means that randomness and probability theories are involved.

Several software applications are commercially available for testing the large number of scenarios associated when taking decisions. Most of these are customised to the specific application field involved. In addition, the methods used result in optimal solutions for the specific problem, which may not be optimal for the situation that actually occurs. Furthermore, the user needs to have a satisfactory scientific background to deal with the investigation of all the possible solutions to the problem posed.

Challenged by these limitations the partners of this project designed the Optimal Decision-Making Application Package for resolving problems and making stochastic decisions. Based on experimental output, this application package employs the Bayesian approach, a classical technique that solves the general likelihood problem subject to general constraints on the parameters involved. This software is especially suited for models with complex constraints on parameters.

The tool is an optimised problem solver that encompasses many special cases of problems found in many branches of the sciences, including mathematics, engineering, economics, ecology, medicine, sociology, physics, operations research, and computer science. In contrast to other similar packages, this one enables its user to set criteria and constraints that are critical for many actual problems in order to minimise the risks involved. Being able to perform risk management, the user is then capable of creating a solution that is optimal over a set of different scenarios.

The package has been designed for PC's running WINDOWS and implements both optimal decision-making algorithms and quasi-optimal hypotheses testing rules that result in significant computer time savings. It is addressed to any user who has little knowledge of mathematics and computer engineering. Highly reliable and simple to use, this new tool is anticipated to contribute substantially both to the theory and the practice of statistical hypotheses testing.

Related information

Record Number: 80450 / Last updated on: 2005-09-18
Domain: IT, Telecommunications