Focusing on the linear case of multiobjective optimization, this chapter provides tractable methods for identifying efficient solutions. The multiple-objective simplex method and interactive techniques such as the Zionts–Wallenius method are presented. The discussion emphasizes interpretability and graphical analysis for two-objective problems. Applications illustrate trade-offs in production planning and resource allocation, such as the Leontief pollution model with multiple objectives and efficiency analysis in a multisectoral economic system, underlining the relevance of the linear structure for multi-criteria decision making. The new approach combining the principles of multiple objectives optimization and DEA—which enables decomposing productivity growth into the movement towards the efficiency frontier and the frontier’s shift as well as to impute the contributions of particular production factors to productivity change over time—is presented.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Multiobjective Linear Programming

  • Mikuláš Luptáčik,
  • Klaus Prettner

摘要

Focusing on the linear case of multiobjective optimization, this chapter provides tractable methods for identifying efficient solutions. The multiple-objective simplex method and interactive techniques such as the Zionts–Wallenius method are presented. The discussion emphasizes interpretability and graphical analysis for two-objective problems. Applications illustrate trade-offs in production planning and resource allocation, such as the Leontief pollution model with multiple objectives and efficiency analysis in a multisectoral economic system, underlining the relevance of the linear structure for multi-criteria decision making. The new approach combining the principles of multiple objectives optimization and DEA—which enables decomposing productivity growth into the movement towards the efficiency frontier and the frontier’s shift as well as to impute the contributions of particular production factors to productivity change over time—is presented.