DSpace Collection:
http://dspace.cityu.edu.hk:80/handle/2031/769
2014-04-07T01:15:18ZSupply chain network design under facility disruptions
http://dspace.cityu.edu.hk:80/handle/2031/6985
Title: Supply chain network design under facility disruptions
Authors: Peng, Peng (彭鵬)
Abstract: Key players in the supply chain, including manufacturers, retailers and distributors,
have realized the value of comprehensive network planning in which
they make detailed plans for constructing new facilities, expanding distribution
networks, partnering with new suppliers, and other important logistics activities.
During the design phase, many parameters in supply chain design problems are
assumed to be fixed. However, the impact of the design decisions spans over a
long horizon, during which many parameters such as costs, demands, and capacities
will fluctuate. Therefore, it will be dangerous to neglect data uncertainties, since a little change in data input may lead to solutions which are far from optimal
in the long run. Possible fluctuations and the reactive strategies have to be
taken into account in order to cope with the uncertain environment.
Many techniques have been derived to deal with optimization problems with
uncertainties, such as sensitivity analysis, stochastic programming methods, robust
optimization and so on. Sensitivity analysis procedures are usually tedious to
implement, while stochastic programming methods often lead to objective functions
that are hard to evaluate. The theoretical framework of robust optimization
has been well developed in recent years. It has received more attention in both
academy and industry. This thesis studies a variety of problems on designing
supply chain networks which are able to achieve well performance in uncertain
environment, with methods derived based on recent development in robust optimization
techniques.
Disruptions represent a form of uncertainty which usually causes capacity loss,
transportation blockage, price inflation and other fluctuations to supply chain networks.
In this thesis, we first study a strategic supply chain management problem
to design reliable networks that perform as well as possible under normal conditions,
while achieving relatively well performance when various forms of disruptions
strike. We present a mixed-integer programming model whose objective is
to minimize the nominal cost (the cost when no disruptions occur) while reducing
the disruption risk by applying the p-robustness criterion (which bounds the
cost in disruption scenarios). We demonstrate the tradeoff between the nominal
cost and system reliability, showing that substantial improvements in reliability
are often possible with minimal increases in cost. We also show that our model produces less conservative solutions than those generated by common robustness
measures. We propose a hybrid metaheuristic algorithm that is based on genetic
algorithms, local improvement, and the shortest augmenting path method
to solve the problem. Numerical tests show that the heuristic greatly outperforms
CPLEX in terms of solution speed, while still delivering excellent solution quality.
The disadvantage of p-robust approach is that it allows less complete description
of the scenario space, since the set of scenarios may grow exponentially large
as the problem size increases. On one hand, only small problems can be solved
due to limited computational power, which makes this model impractical in real
world application, where supply chain networks are usually consisted of hundreds
of facilities. On the other hand, conserving computational power by considering
only a small portion of the total number of scenarios would harm the accuracy of
our results. Traditional stochastic programs which are risk neutral in the sense
that they consider optimization of expected system-wide cost, are also difficult to
solve, since exact evaluation of the expected value is either impossible or prohibitively
expensive. To cope with this computational difficulty, we adopt a Monte
Carlo simulation based method called sample average approximation (SAA), to
solve a stochastic p-robust logistic network design problem in which we minimize
the expected total costs, while enforcing p-robust constraints for each scenario.
SAA approximates the expected objective function of the stochastic problem by
a sample average estimate derived from random samples. It usually results in
MIP counterpart problems and can be then solved by deterministic optimization
techniques. SAA not only approximates, but also produces confidence intervals on the problem's optimal objective values, which makes this method more attractive.
We propose a method based on SAA to solve the stochastic robust model.
Statistical lower and upper bounds are derived as well to evaluate the solution
quality. Numerical test results show that high quality solutions can be obtained
with reasonable computational efforts.
Notes: CityU Call Number: HD38.5 .P465 2011; xvi, 144 leaves : ill. 30 cm.; Thesis (Ph.D.)--City University of Hong Kong, 2011.; Includes bibliographical references (leaves 134-144)2011-01-01T00:00:00ZVehicle and manpower routing problems
http://dspace.cityu.edu.hk:80/handle/2031/6983
Title: Vehicle and manpower routing problems
Authors: Che, Chan Hou (謝振濠)
Abstract: This thesis studies a set of vehicle and manpower routing problems for multitrip/
multiperiod planning, which can be widely applied in practice. Planning
the routes of vehicles, e.g. trucks or ambulances, is usually applied in product
and people transportation, while manpower routing occurs when scheduling field
technicians to perform tasks at customer sites. Making an efficient and practical
routing plan for vehicles and manpower can greatly affect the effectiveness and
service levels of a company.
Most existing literature on vehicle routing problems studies single period routing
(e.g. a day), where the vehicles usually depart and return to the depot only
once within the period, and the objective is to minimize the operational cost.
However, this assumption may not satisfy other important business requirements,
such as labor law regulation. Even when the planning is performed within a single
period, the vehicles may depart and return to the depot multiple times. Therefore,
more
exible routing approaches that consider these factors (such as multiperiod
and multitrip planning) are more likely to be practically applicable.
The contributions of this thesis are threefold. First, we developed some effective
algorithms to resolve real world routing problems for mulitrip and multperiod
planning. Second, we introduce new variants of vehicle and manpower routing problems to the literature. Third, we contributed a set of real-world data to
research community.
The problems investigated in this thesis are all derived from the real world
projects, namely an inspector routing problem in a major retail distributor, a
periodic vehicle routing problem in one of the largest restaurant chains in Hong
Kong, and a non-emergency ambulance transfer service in public hospitals in
Hong Kong.
The first problem studied in this thesis is an inspector scheduling problem for
a major retail distributor. Most existing literature on variations of the vehicle
routing problem assumes that all vehicles are in service within the entire planning
horizon. However, this assumption may not be valid in practice for some applications
due to working time regulations. We formulate the inspector scheduling
problem as a multiperiod vehicle routing problem with profit (mVRPP), where
the goal is to determine routes for a set of vehicles that maximizes the amount
of reward collected from the visited locations, and the vehicles can only travel
during working hours within each period in the planning horizon. Furthermore,
the vehicles are only required to return to the depot at the end of last period.
We propose an effective memetic algorithm with a giant-tour representation to
solve the mVRPP. To efficiently evaluate a chromosome, we develop a greedy
procedure to partition a given giant-tour into individual routes, and prove that
the resultant partition is optimal. We evaluate the effectiveness of our memetic
algorithm with extensive experiments on a set of modified benchmark instances.
The results indicate that our approach generates high-quality solutions that are
reasonably close to the upper bounds and significantly better than the solutions
obtained using heuristics employed by human schedulers.
We next study a periodic vehicle routing problem encountered by a restaurant
chain in Hong Kong. To develop customer relationships and increase service efficiency due to familiarity, some companies prefer their service personnel to
visit regular customers at approximately the same time each day when service
is required. However, this type of service consistency may result in increased
operational cost, especially when the customers' demands
uctuate significantly
day by day. In this paper, we investigate a variant of the periodic vehicle routing
problem with time windows that includes a limited visiting quota constraint
(PVRPTW-LVQ), which requires that any particular customer may be serviced
by at most R different vehicles over the planning horizon. The objective is to
first serve all customers with a minimum number of vehicles, and then reduce
the total distance traveled. We formulate the problem as a mixed-integer linear
program, and also propose a three-stage approach combining several search
techniques for the problem. Extensive computational experiments on benchmark
instances show that the proposed method outperforms previously published approaches
for both the PVRPTW-LVQ and the consistent vehicle routing problem
(ConVRP), which is a related problem where R=1. We also empirically examine
the effects of varying levels of service consistency and demand uncertainty using
our approach, which provides additional insights on the trade-offs between these
two factors in terms of operational cost.
The third problem we investigated is the non-emergency ambulance routing
problem. This problem is derived from the non-emergency ambulance transfer service
(NEATS) in Hong Kong public hospitals. The NEATS provides transportation
service for disabled and elderly patients between hospitals and residences.
The efficiency of the NEATS can greatly affect the operations of the hospital.
We model this problem as a multitrip pickup and delivery problem with time
windows (MT-PDPTW). We devised a fast heuristic to solve the problem, and
performed experiments on various sets of data to evaluate the performance of the
algorithm. We then tested the algorithm on real world data, which showed that our proposed algorithm can solve the problem quickly. We also discussed the
relationship between service level and capacity in the NEATS system.
Notes: CityU Call Number: QA402.6 .C45 2011; xii, 119 leaves : ill. 30 cm.; Thesis (Ph.D.)--City University of Hong Kong, 2011.; Includes bibliographical references (leaves 109-119)2011-01-01T00:00:00ZPipeline and vehicle transportation problems in the petroleum industry
http://dspace.cityu.edu.hk:80/handle/2031/6502
Title: Pipeline and vehicle transportation problems in the petroleum industry
Authors: Zhen, Feng ( 甄峰)
Abstract: In the petroleum industry, petroleum product logistics can be divided into two
phases: first logistics, which is mainly provided through pipeline transportation
or railway, refers to distribution from refineries to oil depots; and second logistics, which is primarily supported by vehicles, pertains to distribution from oil
depots to oil stations. This thesis studies three petroleum product transportation problems faced by transportation practitioners in the petroleum industry:
one stems from first logistics and two from second logistics.
Oil product transportation costs currently account for a proportion of sales
fees in the Chinese petroleum industry that is considerably higher than the average international level. Hence, reducing costs incurred from the transportation of oil products has become a highly important problem for the managers of
Chinese oil companies. This thesis aims to provide a reference for oil companies
for reducing both first and second logistics expenditures. The investigation of
these problems was motivated by actual projects for China National Petroleum
Corporation (CNPC).
For first logistics, a three-phase optimization model for the transportation
of multiple petroleum products using pipelines is described. Through this method, we aim to ensure that all depots are able to satisfy the demand
for each petroleum product while minimizing costs. The first phase involves
solving a mixed integer programming model to create resource allocation plans.
This phase minimizes the number of products transported in each time period.
The second phase uses the output from the first phase and integrates it into a
quadratic mixed integer programming model to create a scheduling plan, which
minimizes pumping costs by selecting the optimal pumping configuration and
flow rate. We employ dynamic programming to increase the e±ciency of the
algorithm, which enables a commercial linear programming solver to address
problem instances of a practical scope. Finally, the third phase post-processes
the solution from the second phase to minimize mixture costs using dynamic
programming. This research was conducted on behalf of CNPC in mainland
China, with findings resulting in annual savings exceeding 1 million Yuan.
For second logistics, we discuss a new practical variant of the vehicle routing problem with time windows (VRPTW), which originated from the regional
transportation planning for oil products at a China National Petroleum Corporation (CNPC) branch in a northwest province of mainland China. Tanker
trucks are scheduled to serve each oil station in multiple periods according
to a recurring and dynamic time window setting. Refilling at an oil depot is
always required after visiting an oil station, so it is safe to assume that the
vehicles are uncapacitated. The problem is formulated into a mixed-integer
programming model and shown to be NP-hard. We found that the mixed-integer programming model is only solvable for very small impractical cases
using exact methods, e.g., branch and cut, which is employed by the state-of-the-art commercial solver ILOG CPLEX. Moreover, due to the floating time
windows imposed on the nodes, traditional local search-based heuristics with node interchange operators are not applicable. Thus, we adapt and propose
an iterative time window partitioning heuristic that discretizes time windows
into multiple time points with dynamic partition widths. Experiments show
good quality solutions can be achieved for problem cases with practical sizes.
In times of uncertainty, transportation demand changes seasonally as the
consumption of oil products fluctuates depending on season. CNPC owns a
limited number of vehicles dedicated to transportation requirements during
regular seasons. During peak seasons, they need to outsource some transportation jobs to third party logistics (3PL) providers because the demand
for oil products (and correspondingly the transportation demand) at this time
is considerably higher. Therefore, the solving of two problems of oil product
transportation from oil depots to oil stations during peak seasons are necessary: first, determine which of the transportation requirements of oil stations
should be outsourced to 3PL providers; second, devise the scheduling plan
that determines which of the oil stations' transportation requirements will be
handled by the vehicles of the petroleum company. This thesis integrates the
combinatorial auction (CA) and vehicle routing problem with time windows
(VRPTW) into a single problem. The problem is formulated into a mixed
integer programming model and shown to be NP-hard. We devise a heuristic
to separate all the stations into two types (depending on whether it is out-sourced to 3PL companies) according to distance. We then obtain an initial
solution by separately solving the CA and VRPTW problems. To improve the
initial solution, we design and test multiple heuristic operators to interactively
solve the CA and VRPTW. Experiments show that good quality solutions are
achieved for problem cases of practical scope.
Notes: CityU Call Number: HE199.5.P4 Z45 2011; xi, 109 leaves : ill. 30 cm.; Thesis (Ph.D.)--City University of Hong Kong, 2011.; Includes bibliographical references (leaves 99-109)2011-01-01T00:00:00ZConditional inference in generalized linear mixed models : model identification and robust estimation
http://dspace.cityu.edu.hk:80/handle/2031/6501
Title: Conditional inference in generalized linear mixed models : model identification and robust estimation
Authors: Yu, Dalei ( 喻達磊)
Abstract: In this thesis, statistical inference problems in generalized linear mixed
models (GLMMs) are considered. In particular, model identification and
robust residual maximum likelihood (REML) estimation for the GLMMs
are studied in detail.
The formulation and estimation for the GLMMs are first reviewed, and the
differences between conditional likelihood based and marginal likelihood
based methods are then discussed. Simulation results indicate that both
methods are promising when the sample size is relatively large. The REML
estimation method is effective in reducing the negative bias in the estimation
of the variance component parameters when the sample size is small.
To address the problem of model selection in the GLMMs, a model identification
instrument based on the conditional Akaike information (cAI) is
developed. In particular, an asymptotically unbiased estimator of the cAI
(denoted as cAICC) is derived as the model selection criterion, which takes
the estimation uncertainty in the variance component parameter into consideration.
The relationship between bias correction and generalized degree
of freedom for GLMMs is also explored. Simulation results show that
the estimator performs well. An adjusted model selection criterion (denoted
as cAICA), which is based on heuristic arguments, is also proposed
as an alternative tool for model identification. Both criteria demonstrate high proportion of correct model identification for GLMMs. Three sets of
real data (i.e. epilepsy seizure count data, polio incidence data and US
strike data) are used to illustrate the proposed model identification methods.
To limit the effect of outliers, a robust version of the REML estimation
for Poisson log-linear mixed model is developed. The method not only
provides robust estimation for the fixed effect and variance component
parameters, but also gives robust prediction of the random effects. Theoretical
and numerical aspects of the estimators are examined. Simulation
results show that the proposed method is effective in limiting the effect
of outliers under different contamination schemes. The epilepsy seizure
count data are used to illustrate the method.
The robust REML estimation method is then extended to the k-component
Poisson mixture model with random effects. The behavior of the estimator
is studied, and the formulae for obtaining the asymptotic covariance matrix
are derived. Simulation study shows that the performance of the proposed
robust REML estimator is comparable with the conventional REML estimator
for regular data, and it outperforms in the presence of outliers. The
urinary tract infections data are taken to demonstrate the proposed robust
estimation method.
Following similar lines of derivations, extensions of the developed methodologies
are possible for a general class of hierarchical generalized linear
models and generalized additive models. These topics are considered as
future research directions.
Notes: CityU Call Number: QA276 .Y8 2011; x, 135 leaves : ill. 30 cm.; Thesis (Ph.D.)--City University of Hong Kong, 2011.; Includes bibliographical references (leaves 101-110)2011-01-01T00:00:00Z