Paper1 : iterative heuristic
order is determined by experiment down-sizing is not useful
doesn't garantee to remove all MIA violation
Paper2 : MS < MIA
Single-row opt
MILP method
strange : left-right subspace
All whitespace filled with fillers->not opt?
single MIA for all Vt
minimize changeVt penalty and displacement distance
For fix macro -> divide the problem->not opt
Paper3 : placement site step?
design utility rate?
why after flipping -> no area overhead?
By 許泓崴
