Watson Center, USA
- ORMA Invited Speaker
"Separation of partition
inequalities and their role in network design"
Partition inequalities are used in network design problems to impose
different types of connectivity constraints. We describe several classes
of these inequalities, we give separation algorithms and discuss their
applications to different Network Design problems.
Oleg Burdakov (Linköping
regression: algorithms and applications"
isotonic regression problem (IR) has important applications in
statistics, operations research and image processing. It can be
formulated as a quadratic programming problem of finding the
n-dimensional vector x that minimizes the Euclidean distance from a
given vector to a cone. The cone is defined by linear constraints
establishes relations between some pairs of components of x in the
form "component number i is less-or-equal to component number j". The
relation between the components can be presented by an acyclic
directed graph. The applied IR problems are often characterized by a
very large value of n. Therefore, the complexity of IR algorithms is
required to rise with n not too rapidly. The IR problem is known to be
of polynomial complexity.
our presentation, we discuss applications of the IR problems and give
an overview of optimization algorithms developed for solving these
Universidad Miguel Hernández:
Solving Mixed 0--1 Stochastic Programs"
present a framework for solving mixed 0-1 multi-stage problems under
uncertainty in the objective function coefficients and the
right-hand-side. A scenario analysis scheme with full recourse is used.
The constraints are modelled by a splitting variable representation via
scenarios. So, a mixed 0-1 model for each scenario is considered plus
the non-anticipativity constraints that equate the so-called common
continuous and 0-1 variables from the same group of scenarios in each
stage. A Branch-and-Fix Coordination approach is presented for
coordinating the selection of the branching Twin Node Families (TNF) and
the branching common variables in the scenario subproblems to be jointly
optimized. We consider Lagrangean Substitution and Decomposition schemes
for bounding purposes at the so-called candidate and integer TNFs. Some
computational experience is reported for different types of problems.
Canada Research Chair in
for the Vehicle Routing Problem: Fifteen Years of Research"
Over the past fifteen
years several powerful metaheuristics have been developed for the
Vehicle Roputing Problem (VRP). The best methods are based on tabu
search, variable neighbourhood search, genetic search, and ant
algorithms. Much progress has been accomplished since the publication of
the first tabu search heuristic for the VRP in 1989. Several methods
have been proposed, but not all have been equally successful. In this
talk I will provide an overview of some of the best algorithmic ideas
proposed over the past fifteen years, and I will also mention some ideas
that did not work so well.
Georgia Institute of
and Optimization Opportunities"
Traditionally companies have focused on improving their own internal
business processes when faced with pressures to operate more efficiently
and more cost effectively. However, a system-wide focus, e.g., a
collaborative focus, opens up cost saving opportunities that are
impossible to achieve with an internal company focus. With the
possibility of sharing and analyzing data through the connectivity
provided by the internet, there has recently been a shift of attention
towards controlling and reducing system wide costs and sharing these
cost savings to increase profitability for all parties involved.
With the availability
of timely and accurate information, new collaborative opportunities
which create mutual benefit by taking advantage of operational synergies
between buyers, sellers, or buyers and sellers are now arising.
Collaborative logistics is viewed by many logistics professionals as the
most promising opportunity for reducing logistics costs and therefore
increasing profitability and economic prosperity.
Probably one of the
most successful applications of collaboration in logistics to date is
vendor managed inventory (although it is not typically presented as an
example of logistics collaboration). In environments where vendor
managed inventory partnerships are in effect, the vendor is allowed to
choose the timing and size of deliveries. In exchange for this freedom,
the vendor agrees to ensure that its customers do not run out of
product. In a more traditional relationship, where customers call in
their orders, large inefficiencies can occur due to the timing of
customers' orders, i.e., high inventory and high distribution costs. By
initiating vendor managed inventory partnerships demand variability is
decreased, reducing inventory holding and
Another, more recent,
successful application of collaboration in logistics is found in the
trucking industry. To execute shipments from different shippers a
carrier often has to reposition its assets, i.e., trucks. Shippers have
no insight in how the interaction between their various shipments
affects a carrier's asset repositioning costs. However, shippers are
implicitly charged for these repositioning costs. No single participant
in the logistics system controls asset repositioning costs, so only
through collaborative logistics initiatives can these costs be
controlled and reduced. Asset repositioning is expensive. A recent
report estimates that 18% of all trucks movements every day are empty.
In a $921 billion U.S. logistics market, the collective loss is
staggering: more than $165 billion.
In this presentation,
we introduce a variety of challenging optimization problems that arise
as a result of these collaborative logistics initiatives and discuss the
potential solution approaches.
Université de Montréal,
Random Number Generation: Overview and Recent Developments"
talk, we first outline a set of design principles for uniform random
(RNGs) used for stochastic simulation. We recall the main requirements
generator (good multidimensional uniformity, high speed, etc.) and
figures of merit for certain classes of linear-type generators. We also
versus statistical testing of RNGs.
are still well alive. As an illustration, we briefly examine those in
Basic, and the Java standard library, and exhibit two very simple
which these RNGs give totally wrong results.
summarize some recent ideas for constructing fast and reliable
They include: (a) combined multiple recursive generators with coe_cients
that are a sum
of a few powers of 2; (b) combined generators whose components are based
linear recurrences modulo 2 (such as Tausworthe, twisted GFSR, etc.);
linear congruential generators with tempering; (d) mixted
linear/nonlinear combined generators.
Practical random number packages with multiple streams and substreams
at the end of the talk.
Several papers on uniform RNGs are available on this speakers’s web