Skip to content
Skip to navigation menu

Objectives

The proposed project seeks to:

  • Establish requirements for intercommunications between a team of problem solvers. The interaction is supported through a Problem Solving Environment (PSE), which enables coupling of various computational and data components from different sites.
  • Investigate means of supporting satisfactory, remote, problem definition – and in particular the relationship between conceptual design and problem specification (design of experiment). An important aspect of this is to enable such an investigation to be carried out in a seamless way – so that the client is unaware that problem optimisation is to be undertaken at a different geographical site. Optimisation is necessary when the problem is complex, and requires analysis of a number of design variables. The methodology used to specify the problem domain is based on workflow. The emphasis in this project is on conceptual design – i.e. the need to create a high-level structure of the system, and identify interaction between entities, which the designer feels are essential for the system to function effectively. Such a conceptual design is then reformed through a series of stages.
  • Establish structures that allow adaptive search software to remotely access system models for function evaluation, especially in the context of search and optimisation algorithms.

Figure 1: Software Interaction

A client will interact with the PSE through a Java based graphical interface, which enables the composition of an “experiment” using the data flow approach. An experiment in this context represents the exploration of various design alternatives, and an understanding of constraints that impact the design. A user will be able to connect together components which represent data services, the Modeller service, the Optimiser service, and visualisation services. These services will be integrated through specialised control objects, which enable the specification of flow-control (such as for-loops, or iterators) and conditionals. The Modeller service has various attributes that can be specified through a scripting language. The execution of the services is controlled through a Web Services Description Language (WSDL) interface. Similarly, the Optimiser can be managed using a configuration file – which may either be predefined or altered by the designer – and also has a WSDL interface. Multiple instances of the Modeller and Optimiser can co-exist, each distinguished by a unique service identifier.