Multiobjective programming

INTRODUCTION

Related to linear, integer and non-linear programming, multiobjective programming is concerned with the extensions to theory and practice that enable us to address mathematical programming problems with more than one objective function.

In single objective programming, we often settle on a single objective such as maximize profit or minimize cost. However, in many if not most real world problems, we may find that we are in an environment of multiple conflicting criteria. To illustrate problems that may be more adequately modeled with multiple objectives, we have:

• Oil Refinery Scheduling

  • min {cost}

  • min {imported crude}

  • min {high sulfur crude}

  • min {deviations from demand slate}

• Production Planning

  • max {total net revenue}

  • max {minimum net revenue in any period}

  • min {backorders}

  • min {overtime}

  • min {finished goods inventory}

• Forest Management

  • max {timber production}

  • max {visitor days of recreation}

  • max {wildlife habitat}

  • min {overdeviations from budget}

  ...