New economic and environmental challenges have recently led to a growing interest in modern mathematical optimization and modeling techniques in order to improve the froth flotation performance. An important issue is that most of corresponding mathematical optimization models are non-convex and involve multiple criteria; hence, conventional convex optimization methods cannot guarantee finding global (Pareto) optimal solutions. We propose a deterministic biobjective mathematical programming framework, combined with experimental design and regression analysis, to optimizing flotation performance and determining the optimal operating conditions that meet specific needs. The framework aims at maximizing concentrate grade and recovery and is based on some well-known and advanced approaches including an epsilon-constraint method, DC programming, an exact penalty method, and the special global search strategy. To demonstrate the effectiveness of the proposed approach, we present a case study for the rougher flotation process of copper-molybdenum ores performed at the Erdenet Mining Corporation Mineral Processing Plant (Mongolia). (C) 2017 Elsevier Ltd. All rights reserved.