Welcome
Quite a
few papers are currently used or implemented in several companies, and for this
reason, only selected publications will be posted here. These
papers are organized according to the application fields as follows.
Patents
·
System and Method for Optimizing Equipment Schedules,
U.S. Provisional Application No.
60/507,504
Paper Under Review
Working Paper to Be submitted in Three Months
·
Aleman,
R., X. Zhang and R. Hill, “An Adaptive Memory Algorithm for the Split Delivery
Vehicle Routing Problem,” Journal of
Heuristics, Revision Submitted.(2008)
·
Gu, Q., X. Zhang, S. Narayanan (2008). “A New LP Based Neighborhood Search Heuristic to Solve
Knapsack Problems,” Working Paper,
Department of Biomedical, Industrial and Human Factors Engineering,
Write State University, Dayton, OH 2008, (to be submitted)
·
Zhang, X., “A
Tabu Search Algorithm for the Clustering Problem,” Working Paper, Department of Biomedical, Industrial
and Human Factors Engineering, Wright
State University, Dayton, OH 2005, (to be submitted)
·
Zhang, X.,
“Solving Multi-Period Clustering of Customers in a Supply Chain Optimization
Using Tabu Search,” Working Paper, Department
of Biomedical, Industrial and Human Factors Engineering, Wright State University, Dayton, OH 2005, (to be
submitted)
Reports
·
Zhang, X., G.
Song, G. Yu and J. F. Bard, “Model and Solution approaches to Solve Large Scale
Crew Recovery Problems during Airline Irregular Operations,” Working Paper,
Department of Biomedical, Industrial and Human Factors Engineering, Write State
University, Dayton, OH 2005
Academic
Research Summary:
My research is in the field of mathematical
programming and optimization, especially linear and integer programming and its
applications in Manufacturing and Service
Operations, Logistics, Transportation and Engineering Optimization. Some specific projects that I have worked
or is working on are listed as follows.
U.S.P.S. is the second largest
employer in the U.S. with over 800,000 personnel. More than 1/3 of these employees work at
around 275 mail processing centers nationwide.
The central problem here is how to run these facilities efficiently with
minimum labor cost. The problem is
modeled as a mixed flow shop and job shop problem with multiple criteria and is
solved sequentially. The core of this
optimization scheme is the inclusion of a surrogate constraint in the form of
worker shifts that must cover daily labor costs. The inclusion of this
constraint is vital and provides a way to combine equipment and labor
scheduling. Several computational
improvements, including pre-processing, the use of target solution and
decomposition have been elegantly combined to find high quality solutions
efficiently. A decision support system
is developed and is now being tested at several facilities. It is estimated that annual savings on the
order of $1.6 million can be achieved per facility (there are 275 facilities
nationwide). The system is expected to
be fully deployed in the next three years.
2)
Crew Management in Airline Irregular Operations (One
of the Key Developers)
An airline schedule
rarely operates as planned due to various disruptions. These disruptions, if not managed quickly and
correctly, could propagate and amplify through the system. In this project, we investigated how to
quickly bring the crew back to their published schedule. The basic problem was modelled as a set
covering with side constraints and variations were investigated to allow robust
schedule. Several algorithms including
simulated annealing and Tabu search heuristics, customized branching together
adaptive improving techniques were developed to achieve faster
convergence. I assisted with the implementation in Northwest Airlines. The project was later awarded
the Franz Edelman Prize from INFORMS in 2002.
3)
Advertising Allocation for TV Networks (Chief
Developer)
Television networks depend on
selling a limited number of advertisement slots in their entertainment programs
to advertisers to sustain their operations and to deliver programs to homes
free of charge. The problem here is how
to efficiently allocate these slots to achieve maximum profits and to meet the
advertisers’ demographic coverage requirements.
A two-step hierarchical approach, composed of a winner selection problem
and a pod adjustment problem, was designed.
The winner selection problem was then solved using an efficient column
generation-based algorithm. The
methodology is applicable to similar advertising allocation problems in other
areas, such as cable networks and radio stations.
4)
Design and Optimization of LFSR Based VLSI Testing
Structure (Chief Developer)
Testing of digital circuits includes performing test
pattern generation and output analysis and is a major portion of the effort in
the design, production and usage. One of
a major scheme in VLSI testing is to perform test pattern generation and output
response analysis through built-in hardware such as Linear feedback shift
registers (LFSR). This research
investigates the mathematical model and methodologies to the structure design
and optimization of 2D LFSR based VLSI Built-in Self-Test. Preliminary results show that hardware
reduction nearly 80% was achieved -- the optimal size of the testing hardware
is only 1/6 as that used reported in the literature. This research, coupled with the study of test
relaxation technique, will provide a novel and direct route to the design of
VLSI system with high fault coverage and is expected to reduce testing hardware
by orders of magnitudes.
5)
Novel Neighborhood Search for Integer Program with
Multiple Solutions (Chief Developer)
Integer programs arise in a variety of engineering
and management areas in manufacturing, logistics and supply chain. Though there
have been tremendous advances in computational power and discrete optimization
solution techniques, integer program remains one of the hardest problems to
solve. Most advanced algorithms require deep knowl
6)
Supply Chain Network Planning, Inventory and
Warehouse Management (Chief Developer)
This research includes supply chain network planning,
inventory control, warehouse management, and manufacturing strategies and the
design of state-of-the-art assembly-to-order systems. One such system, an
assembly to order system have saved companies more than millions in the raw
materials cost, reduce lead time in getting products to customer,
Industrial Project Summary and Impacts:
A partial
list of past projects in supply chain and logistics transportation and
logistics, inventory control, and warehouse operations is listed below.
·
Logistics and
Transportation in Airlines -- Implemented
in Continental and Northwest Airlines, more
than $50M reported for a major airline in 2001.
·
USPS
Equipment and Workforce Scheduling -- Implemented in 275 facilities nationwide with $400M savings reported.
·
Forecasting, Production Planning, Manufacturing, and
Inventory – Rittal Corp : More than $1M savings out of
a $12M business department.
·
Material
Management, Warehouse Layout and Management – Rittal Corp: The warehouse being significantly
consolidated – 50% of its original size; Picking efficiency has greatly
increased by 200%, picking time was reduced from 29 minutes to 8 minutes; Labor
and overtime cost reduced by 20%.
·
Inventory
Management, Warehouse and Lean Manufacturing – Rittal Corp and Peco II cooperation. More than $3M, 25% reduction in
inventory is reported in Peco II from 11/2007 until 03/2008, another $3M is
expected in the next 6 Months.
·
Procumbent
Management and Combinatorial Auction– NBC & Sears Logistics and David Joseph
Company: $500,000 expected out of $5M, 10% reduction in transportation cost.
·
Facility
Location in Reverse Logistics – Hilti Corp: The exact location and inventory for
Hilti Corporation’s 11 warehouses and distribution centers, currently adopted
by Hilti Corp