Multiobjetive optimization co-evolution for the job shop scheduling problem

Abstract

A job shop is a facility that produces goods according to specified production plans under several domain-dependent constraints. Job Shop Scheduling (JSS) attempts to provide optimal schedules. Cornmon variables to optimize are total completion time (makespan), machine idleness, lateness and total weighted completion time. According to this variables different objectives can be devised. Multiobjective optimization, also known as vector-valued eriteria or multieriteria optimization, have long been used in many applieation areas where a problem involve multiple objeetives, often eonflieting, to be met or optimized. Co-evolution, as an extended evolutive model, ean be applied to solve multieriteria optimization for the JSS problem using a plain aggregative approach. This presentation will show the design, implementations and results of a co-evolutive approaeh solving a multiobjeetive optimization problem involving the makespan, maehine idleness and total weighted eompletion time as eriteria to be optimized.