I was thinking that when researching weapons, defense, and engines, that the improved designs should constitute a separate branch rather than just flowing into the next 'innovative' tech.
Something like this:

New Propulsion Techniques
|
V
Ion Drive
|
V
Impulse Drive -> Impulse drive 2 -> Impulse Drive 3
|
V
Warp Drive -> Warp Drive 2 -> Warp Drive 3 -> Warp Drive 4 -> Warp Drive 5
|
V
Hyperwarp Drive -> Hyperwarp Drive 2 -> Hyperwarp Drive 3

-------
The idea is that improving an existing tech is essentially miniaturization, and doesn't really help one to come up with a breakthrough revolution in propulsion (or whatever); that requires researching totally new concepts. It is, however, significantly easier to improve an already existing design than it is to create a new one from scratch, so research down an 'improvement' line is relatively quick and cheap, whereas researching the 'innovative' idea would take much, much longer, like several multiples of the time it takes to completely improve the previous tech, but you end up with a significant advantage. This would cause players to consider some short vs long term strategies concerning research.

Comments?