Linear Modeling Optimization for Workload Assignments

Abstract

Company XYZ (XYZ) operates a government contract in south Kansas City Missouri. XYZ is most notably engaged in the manufacturing of high-precision components for national security. XYZ operations are heavily dependent on complex, custom-built test equipment which is spread throughout their 3.1 million square foot facility. XYZ has occupied their current space for over 60 years, but as the infrastructure has aged and the workload requirements have changed dramatically, they are now in the midst of relocating to a new site in Kansas City that allows for today’s highly flexible, rapid response manufacturing needs. The massive XYZ relocation has been phased into 142 unique individual moves that will take over 18 months to complete. The manufacturing floor at XYZ holds over 2,000 unique pieces of test equipment. An entire work group at XYZ is dedicated to the routine calibration and maintenance of this fleet of test equipment. In addition to this routine work that supports production, all test equipment must be “close loop calibrated” prior to relocation to the new facility to ensure product quality levels are being met. These 2,000 pieces of equipment belong to any one of the 142 move phases, but the volume of test equipment in each move phase is highly dependent on the technical area being relocated in a given phase. In February 2014, Test Equipment Calibration & Maintenance (TECM) will be tasked with completing close loop calibrations for the Assembly and Electrical Fabrication (AEF) area of XYZ that includes 16 different technology areas. This specific move phase holds 191 unique pieces of test equipment, and easily dwarfs all other move phases that have calibrated test equipment. Only 10 technicians in TECM are cross-trained for calibration work within any of the given 16 technology areas, which severely limits TECM workload capacity. After individual close loop calibration job assignments were made, initial capacity planning indicated that TECM did not have the resources needed to complete the assigned work in the AEF move phase on time. By creating a linear optimization model that included 50 decision variables and 26 constraints, the researcher has identified key cross-training needs and specific workload assignments that align individual technician efficiencies within specific technology areas to successfully deliver all AEF calibration needs within the relocation timeframe. Based on individual technician efficiencies, the output of the final linear optimization model yields the exact number of work hours each technician needs to work in a given technology area during the AEF timeframe. Technician availability and technology area demands are used as constraints to ensure all other technician commitments and all AEF technology area work is completed while ensuring AEF close loop calibration requirements are met within the identified timeframe.