Next: A View on Concurrent Up: Greedy Algorithms and the Previous: References Contents
I think this is potentially a really good choice of topic, as it has specific aspects (e.g. making change with current UK currency) but admits wide and subtle extensions. This allows you to begin with a relatively simple model and work from there. Certainly good to be able to play around with the denominations of coins available, and use this as a way of exposing the principles behind the greedy algorithm. Might also be good to (at least in passing) look at alternative issues that your model-building might also be adapted to address. For instance, in what circumstances would the greedy algorithm be inappropriate (if that makes sense!) - maybe if some coins were worth more than their face value in combination. Perhaps could ask other questions (like 'how do I give away as many coins as possible in making up a specific amount?'). The design of coins for blind people is another factor that impinges on the design of currency that could be worth thinking about.
Perhaps the theme you are developing concerning the empirical roots of knowledge would be well-illustrated by showing how your model could (amongst other things of course, as is usual for an EM model) be the basis for a practical exercise for children that leads them through the experiences that inform the greedy algorithm.
It would be really good to be able to make connections with other greedy algorithms through constructing links between different representations of problems (cf. the alternative representations for monotone boolean functions investigated in fdl4Beynon2002).
As they stand, your references are a bit on the vague side. I don't know the philsophical stance of Curd and Cover (perhaps they are reviewing many stances in any event) but good if you can to get a closer look at the thinking of David Gooding on the theme of Philosophy of Science (he is quite critical of what he terms as 'received philosophies of science'). Specific references to educational software and studies relating to coins would be good.
Other models to look at might be fractionsCronick2002 and a companion model that was also a 3rd year project of the same vintage.