Optimizing FIR pulse-shaping filter

I discovered this paper (pdf) that tackles a general problem that is not dissimilar to what's asked in the contest. Could elements of the discussion in this paper be used to approach the optimization problem given?