Coordinated Scheduling of Customer Orders for Quick Response
Reza Ahmadi,
1
Uttarayan Bagchi,
2
Thomas A. Roemer
3
1
Anderson School of Management, University of California at Los Angeles, Los Angeles, California 90095
2
Department of Management, CBA 4.202, The University of Texas, Austin, Texas 78712
3
Sloan School of Management, Massachusetts Institute of Technology, Cambridge, Massachusetts 02142
Received 3 December 2002; revised 12 April 2005; accepted 12 April 2005
DOI 10.1002/nav.20092
Published online 13 June 2005 in Wiley InterScience (www.interscience.wiley.com).
Abstract: The scheduling problem addressed in this paper concerns a manufacturer who produces a variety of product types and
operates in a make-to-order environment. Each customer order consists of known quantities of the different product types, and
must be delivered as a single shipment. Periodically the manufacturer schedules the accumulated and unscheduled customer
orders. Instances of this problem occur across industries in manufacturing as well as in service environments. In this paper we
show that the problem of minimizing the weighted sum of customer order delivery times is unary NP-hard. We characterize the
optimal schedule, solve several special cases of the problem, derive tight lower bounds, and propose several heuristic solutions.
We report the results of a set of computational experiments to evaluate the lower bounding procedures and the heuristics, and to
determine optimal solutions. © 2005 Wiley Periodicals, Inc. Naval Research Logistics 52: 493–512, 2005.
Keywords: sequencing; customer orders; algorithms and complexity; bounding procedure
1. INTRODUCTION
This paper discusses a sequencing problem that occurs in
many manufacturing and service environments. Consider,
for example, a product development team whose members
independently develop modules for a number of different
products. A product design is only completed once all
modules have been designed. A natural performance crite-
rion for the development team is the average completion
time of all product designs, or, in the presence of ranked
importance of the designs, the weighted average completion
time. A team of accountants auditing distinct parts of dif-
ferent companies is another example. The customer, the
audited company, receives a final report only after all ac-
countants have finished their work. Again, the (weighted)
average completion time, that is, the time until the company
receives the final report, is a crucial performance measure.
In manufacturing, this problem arises in many assembly
situations. Assembly can only proceed if all parts for as-
sembly are available. Thus, completion time is largely de-
termined by the time that the last component is manufac-
tured. Under 0-inventory policies or mass customization
schemes where each component is unique, the time until the
customer receives the assembled product is then mainly
determined by the manufacturing schedule of the compo-
nents.
Specifically, this research is motivated by a manufacturer
that produces semifinished lenses and competes in the
global market. The lenses are sold to professional lens
finishing labs, and some large optometrist stores. The man-
ufacturer produces three types of plastic lenses: CR-39,
Polycarbonate, and Spectralite. The CR-39 is the industry
standard for plastic lenses. The Polycarbonate lenses are
made of higher index materials than CR-39, which give
longer durability to this type of lens. The Spectralite lenses
are made of higher index materials but do not require as
much curing as the Polycarbonate lenses. Due to the nature
of processing requirements, each type of plastic lens is
produced on a dedicated production line. The manufacturer
produces lenses based on confirmed customer orders. Each
customer order consists of different quantities of the three
lens types. Different parts of the customer order, after
completion on the different production lines, are sent to the
Correspondence to: T.A. Roemer (troemer@mit.edu); R. Ah-
madi (rahmadi@anderson.ucla.edu); U. Bagchi (uttarayan.bagchi@
mccombs.utexas.edu)
© 2005 Wiley Periodicals, Inc.