First, get rid of the home/road multipliers (we'll use them in a minute). Next, calculate the RPI as it was done until ~2005 (25% your winning percentage, 50% your opponent WP, and 25% opponents' opponents).
EXAMPLE ONLY:
Now, for every win you get below 100, you get nothing. For every loss below 150, you lose 1 point (don't want to penalize teams in decent conferences who don't sweep #125, for example). For every win above 100, you get 2 points, and for every loss you lose nothing.
For a 75 - 100 win, add a point.
For a 50 - 75 win, add another point.
25 - 50, add a point.
11 - 25, add a point.
0 - 10, add a point.
Separate all points into home and road. Multiply home by 0.8, and road by 1.2.
Add up all the points.
Now, test the calculations against the NCAA tournament field for the past ~10 years. Adjust the point totals and multipliers until it makes sense. Once you've got a formula that replicates the field fairly well, use it. You'll never have the problem of scheduling again. Zero risk for playing tough Tuesday games, zero punishment for playing easy ones. Northern teams will come down south in droves because they need the points, etc.
Now, you'll get some really good teams that may overschedule, wind up 28-28, etc. and their RPI may be around 100. This will hurt the point totals of the teams that played them. But the good news for that .500 team is that if they have 5 wins over top 10 teams, they might make the tournament. I have no problem with that, as long as the AQ system remains.