A MultiObjective Decision Support System for Workforce Training

1
A Multi-Objective Decision Support System for
Workforce Training
Project CDP07-UL
Project Team Principal Investigator Dr. Gerald
Evans Co-Principal Investigator Dr. Gail
DePuy Graduate Assistant Brandon Elmes

• Project Activities
• Literature review on multi-objective optimization
• Determine appropriate training times and costs
• Code multi-objective heuristic using VBA in Excel
• Present and validate the DSS through interaction
  with NSWC

• Project Objectives
• Extension of previous project which considered
  only one objective function
• Determine the preferred worker-task assignment
  according to the DM
• Worker-task assignment leads to a training
  schedule for workers

• Project Motivation
• Naval Surface Warfare Center, Crane, IN
• Third largest Navy installation in the world
• .5 M for each refurbished engine
• Workers are undertrained to perform tasks
• Newer and newer Navy technology makes it hard for
  employees to stay up to date

Research Methodology

• Model Worker-Task Assignment Problem
• Multi-objective decision support system
• Z1 Minimize Training Cost
• Z2 Minimize Skill Gap between current skill
  levels and skill level required by tasks
• Z3 Minimize the Number of Workers who do not get
  Preferred Training
• Constraints on Worker-Task Assignments
• Worker Capacities
• Retain Current Workforce (all workers assigned at
  least one task)

• Heuristic to Solve Formulation
• In a previous project, Meta-RaPS was used to
  solve Z1Minimize Training Cost
• Meta-RaPS is used to solve the other objective
  functions for Phase 1Generate Initial Solutions
• In Phase 2, the DM compromises certain objectives
  for the benefit of others
• Phase 2 is iterative and continues until the DM
  is satisfied with the solution
• This method does not use subjective weights on
  objective functions