Next: Empirical Modelling for Learning Up: Modelling a Computational Grid Previous: References Contents
There is a bold objective here, and you haven't admitted the possibility of compromise (such as 'assessing the feasibility', or 'serving an educational purpose' e.g. in illustrating how economic models might be applied to Grid computing). Certainly, it would be great if you can make significant progress towards achieving your goals for realistic applications, but this is presumably only conceivable if you take care to make any visualisation and processor / resource models etc minimal. I like the idea of incorporating research from an outside source, and reading and applying the paper from Monash will be a good exercise. It would also already be quite enough I feel for you to construct models that can show how different economic models apply in principle, and/or to discuss in detail how just one economic model can be applied. It might also be worth saying something about the theoretical complexity of the kind of optimisation that you're studying here - presumably there are known NP-completeness results in this area concerning resource allocation, for instance. Quoting results of this kind will give more force to any positive results that you can demonstrate, and might temper claims such as "A simulation environment would therefore enable the configurations which provide the highest overall throughput and achieve the best load balancing to be determined".
Your writing is very accomplished, but perhaps a little dense. You might consider whether more paragraphs would help to make your points and arguments stand out more clearly.
Not sure about the weighting - it sounds as if 'an entirely new model' should merit more that 30% of the assignment mark - is this what you intend? It is certainly true that there is plenty to write about here though, in terms of challenging issues to be put in perspective. It will be important to clarify where EM is important here, for fear that the paper degenerates into a discussion of Grid computing issues pure and simple, so worth devoting some serious attention in your account to explicit ways in which you think agency, dependency and observation relate to the Grid.