“…Commonly, these optimization-based adaptation approaches treat layout adaptation as a single-objective optimization problem, combining multiple objectives for multiple elements into a single cost to be minimized or a single utility score to be maximized and returning a single solution. Most frequently, this global criterion takes the form of a static weighted sum (e.g., [3,11,16,20,27,32,41]), which can be optimized using commercial solvers for linear programs (e.g., [11,27]) or using approximate techniques like simulated annealing (e.g., [3]). In contrast to these previous efforts, our approach treats layout adaptation as an online multi-objective optimization problem that returns a set of optimal adaptations.…”