Essays on Workforce Scheduling and Cross-Training Decisions with Heterogeneous Learning and Forgetting

Open Access
Author:
Kim, Sungsu
Graduate Program:
Industrial Engineering
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
November 19, 2010
Committee Members:
  • David Arthur Nembhard, Dissertation Advisor
  • David Arthur Nembhard, Committee Chair
  • Paul Griffin, Committee Member
  • Ling Rothrock, Committee Member
  • Spiro E Stefanou, Committee Member
Keywords:
  • Workforce Scheduling
  • Cross Training
  • Workforce Flexibility
  • Learning and Forgetting
  • Productivity
  • Staffing Level
  • Rule Mining
  • Robust Optimization
Abstract:
This dissertation aims at modeling workforce scheduling and cross-training decisions in the presence of heterogeneous individual learning and forgetting characteristics, in which worker productivity changes dynamically over time depending on the previous experience level and assignments. The primary objective of this research is to derive managerial implications in different objectives through appropriate decisions regarding multi-skilled worker allocation. Since a skilled workforce has become an increasingly expensive resource for organizations, determining minimum staffing levels to meet production requirements is critical to competitiveness. This dissertation investigates the effects of several exogenous and controllable factors on minimum staffing levels in parallel dual resource constrained (DRC) systems. The study examines factors of worker selection policies, task heterogeneity, individual cross-training levels, time schedule granularity, and production requirements. The research explores two cases of parallel systems in which a system consists of an equal number of workers and tasks and another with more tasks than workers. In addition, the dissertation presents an association rule mining based framework for workforce scheduling to assist managers with robust real-time assignment decisions. The study explores a parallel production system that meets specified production requirements over a fixed time horizon with the minimum workforce resources based on the number of worker-periods assigned. Three managerial policies are considered including, setting a maximum allowable individual cross-training level, balancing workload among workers, and an unconstrained policy. The research proposes the use of several schedule attributes to quantify key aspects of optimized schedules that may in turn, aid in determining robust assignment rules, and the development of better cross-training policies. Current results indicate that the proposed approach is effective at identifying important rules, many of which add to our knowledge of useful workforce scheduling strategies. Besides, the dissertation examines analytically how heterogeneity in each of the four parameters of the Exponential L/F model affects system productivity. The dissertation also introduces several linearization techniques to reformulate non-convex mixed integer nonlinear (MINLP) workforce assignment models to mixed integer (MIP) models. Through the linearization processes, the resulting model becomes entirely linear, so we can utilize commercially available MIP solvers which is capable of solving problems an order of magnitude larger than typical MINLP solvers can tackle.