DSpace Program:http://hdl.handle.net/123456789/862020-06-04T08:23:37Z2020-06-04T08:23:37ZA Class of Symmetric Distributions with Bell-shaped Densities on Finite IntervalsAbrar Bader AL-Mukhaizeemhttp://hdl.handle.net/123456789/9882019-11-18T08:53:03Z2019-01-01T00:00:00ZTitle: A Class of Symmetric Distributions with Bell-shaped Densities on Finite Intervals
Authors: Abrar Bader AL-Mukhaizeem
Abstract: In this thesis, we introduce and study certain symmetric classes of continuous random variables with bell shaped density functions deﬁned on ﬁnite intervals. Speciﬁcally, we study the random variables, deﬁned over the interval (0,1), that are of the form X = W +ΘV, where the random vector (W,V) follows a Dirichlet distribution of order k = 3, with parameters (α,β,α), and Θ is a continuous random variable with a symmetric distribution deﬁned over (0,1). We investigate the properties of the joint probability distribution function of X, Θ and the marginal distribution function of X. We investigate how diﬀerent probability distributions for Θ aﬀect the distribution of X. Speciﬁcally, we exam the case when Θ follows a Beta(γ,γ) distribution and the case when Θ follows a triangular distribution. For both distributions, we determine the marginal distribution of X analytically, and investigate estimation procedures for the distribution parameters using both simulated and real data.2019-01-01T00:00:00ZApproximate Approaches for Solving the Vehicle Routing Problem with Multiple Time WindowsAmal Mahmoud Matarhttp://hdl.handle.net/123456789/7642019-04-04T08:51:40Z2018-01-01T00:00:00ZTitle: Approximate Approaches for Solving the Vehicle Routing Problem with Multiple Time Windows
Authors: Amal Mahmoud Matar
Abstract: Faced with increasing competition and wishing to remain competitive, distribution
companies are pushed to reduce their logistics’ expenses. Their obligation to protect the
environment, reduce pollution, improve drivers’ work conditions, and deal with higher
fuel prices has geared them toward improving their vehicle routing systems. In this
context, this research studies the vehicle routing problem with multiple time windows
(VRPMTW), where each client specifies more than one time-window for the actual
delivery time. It models the problem as a constraint program and compares it to a
simplified version of an existing mixed integer program. Because VRPMTW is NP-hard,
solving it via an exact method is time consuming and without any guarantee of
convergence to optimality within a realistic runtime. Hence, most proposed methods are
approximate or heuristic. Although a heuristic does not necessarily identify the optimum,
it gives good practical results. This thesis addresses VRPMTW via a particle swarm
optimization algorithm (PSO), and tests its performance on benchmark instances from the
literature. It further hybridizes PSO with a variable neighborhood search (VNS) that
enhances the quality of each particle of the swarm. Finally, it provides computational
evidence of the utility of this hybridization in enhancing the quality of the solutions, but
recommends opting for a pure VNS.2018-01-01T00:00:00ZOn Some Laplace-Type Discrete DistributionsNourah Buhamrahttp://hdl.handle.net/123456789/7632019-04-04T08:51:33Z2018-01-01T00:00:00ZTitle: On Some Laplace-Type Discrete Distributions
Authors: Nourah Buhamra
Abstract: The Laplace distribution (LD) has many applications in Engineering, Finance,
Economics and Astronomy. Several representations of the continuous Laplace distribution
are given in Kotz et al. (2001). Aly (2018) proposed and studied a unified
approach for developing Laplace-type distributions using a random sign transformation
(RST) and a random sign mixture transformation (RSMT). Recently, some
discrete Laplace-type distributions have been developed and studied in the literature.
These distributions are essentially obtained as the difference of two independent
integer-valued random variables. This thesis examines a number of discrete
Laplace-type distributions developed using the RST and the RSMT of the Geometric
and the Poisson distributions. Properties of these distributions are derived.
Moreover, using two real life of data sets, the proposed distributions are compared
with their available counterparts in the literature.2018-01-01T00:00:00ZModelling SF-6D Health State Preference Data using BayesianWadha Ahmad Jassim Shumaishttp://hdl.handle.net/123456789/7582019-04-01T08:51:37Z2018-01-01T00:00:00ZTitle: Modelling SF-6D Health State Preference Data using Bayesian
Authors: Wadha Ahmad Jassim Shumais
Abstract: The thesis presents an approach to modelling SF-6D health states
preference data. It provides a new approach to estimating health state values
data using Bayesian methods. The data set is the UK 6-dimensional short form
health survey (SF-6D) valuation study, which is a generic preference-based
measure of health derivative from the 36-item short form health survey (SF-36).
A sample of 249 health states defined by the SF-6D was valued by a
representative sample of 611 members of the UK general population, using the
standard gamble (SG). The thesis presents the results from applying two random
effect models to the data using a Bayesian approach; one with constant variance
and another with variable variance, and then comparing these results to the
original results of the random effect model estimated using a classical approach.
The thesis also investigates these results for the future implementation of the SF-
6D and further work in the field.2018-01-01T00:00:00Z