Second Experiment Done

Second experiment is now done and dusted as well. The simulation has been updated to remove an entity within the biological system in question. This removal has been introduced via a parameter-driven approach, therefore the initial code relating to the entity in question is still there, but is only called when the correct parameter flag is fed in to the simulator.

Again, results are appreciably different from baseline, and again have been found to be statistically significant through use of an A-Test.

As discussed previously, the remit set was for me to perform 3 updates to simulator functionality, however due to the issues encountered regarding hardware, there is simply not enough time to perform this last update and still generate a quality project report.

I have therefore gained agreement from my supervisor that the 2 sets of experiments performed will be sufficient for the project, and that writing up the project report is now the main priority. There is now 4 weeks until we depart for San Diego, and therefore 4 weeks for me to complete the project - again, I'm sure you all remember that the firm deadline for project report submission is 31st August, but with the week away in the US my actual deadline for finishing the report moved forward by a week.

No Works of Art Required

Out there, in a galaxy far, far away, Project Managers create schedules which are not only objects (works of art as I used to call them), but also are effortlessly stuck to with no deviations from baseline - no scope creep, no unforeseen impacts, no resource constraints, etc, etc.

Now back to reality!

I always used to appreciate the benefits of re-planning projects at pre-defined periods within projects (e.g. the end of a phase), but am now coming around to the concept of iterative scheduling. In a traditional project plan (i.e. the initial schedule on this project), major pieces of work are scoped out in detail and then performed in a predetermined sequence - typically work on stages does not commence until previous stages (sometimes with predecessor tasks) are complete. Iterative scheduling is characterised by less rigid planning and much smaller chunks of work. The key here however is that these chunks of work are only decided upon after previous chunks have been completed.

Javed Iqbal PMP states that "the technique is based on the fact that detailed, accurate planning can only be achieved for tasks that will occur in the near future". Furthermore, "since accurate information is not available for the tasks that will occur in the distant future, their planning will, by definition, be inaccurate".

So, no more works of art need be produced which would not look out of place in the Tate Modern Art Gallery!

First Experiment Done

First experiment is now done and dusted. The simulation has been updated to refine the functionality of one entity/object within the biological system in question.

Results are appreciably different from baseline, and in fact have been found to be statistically significant through use of an A-Test - calculates probability that variables within sample A are stochastically smaller than sample B.

It's now time to move on to design, implementation and simulation of experiment 2.

Time to Really Motor

With a little help from friends, I now have 2 viable clusters from which to work from. The Biology cluster has a Sun Grid Engine installed to automatically manage my experiments, and I've been given access to a script which will perform the same task on the external cluster. This is fantastic news.

First step is to repeat experiments previously performed by the PhD Student who I'm woking with, then it's on to designing code updates, implementing the updates and running new experiments.

It is now time to really motor with this project.