I've been playing around a bit with the new missile design rules.
The basic missile design problem used to be optimizing to hit rate by varying the relative weight of engine and agility points, given a desired range, missile size and warhead size. With the new missile engine system, this is complicated by the fact that there are now a couple of new (and non-trivial) tradeoffs between engine power, agility and fuel use: A larger engine gives more than proportionally greater speed, because you can step up the power boost or reduce the fuel commitment. Conversely, agility can be traded for fuel, which can then be traded for a power boost (given fixed engagement range).
I'll spare you the two pages of algebra, and jump straight to the conclusions: If I do not constrain the agility and power boost values, the optimal configuration assigns negative agility values for all valid missile parameters. This means that the optimum is an edge case - agility is zero, the power boost is minimized, or the power boost is maximized (because all the functions here are nice, well-behaved, continuously differentiable functions.
A few numerical examples suffice to establish that the relevant edge case is not when power boost is minimized: Missile fuel budgets, for realistic fire control ranges, are comparatively trivial relative to engine sizes, so the gains from trading fuel for engine are naturally limited. Nor is it when Agility is zero, because this gives desired power boosts which are well beyond the top end of the tech tree. Which means that the relevant edge case is the one where engine boost is maximized.
Surprisingly, deriving an exact solution for this edge case turns out to be a non-trivial problem, but since fuel takes up a comparatively small part of any missile that you want to actually hit things, you can probably just use the old equation of [engine size] = [size - payload]/2 + 5*[size]/[agility per MSP], and then manually test whether rounding engine size up or down gives you the superior design.
Please note that the usual caveat applies: This optimizes to hit chances - in shipkillers, you typically want to trade some to hit chance for speed to get through point defense, and in busses, you don't want Agility at all, since they're not supposed to hit anything.