\documentclass[12pt]{article}

\newenvironment{tightenumerate}{\begin{list}{--~}{
  \topsep=0.3ex \itemsep=0.3ex \labelsep=1em \parsep=0em
  \listparindent=0em \itemindent=0em
  \settowidth{\labelwidth}{--~} \leftmargin=3.5em
}}{\end{list}}

\textwidth=6.5in \textheight=8.5in \oddsidemargin=0in 
\evensidemargin=\oddsidemargin \topmargin=0in

\begin{document}
\begin{center}
{\LARGE High Performance Optimization:} \\ \medskip {\large 
Theory, Algorithm Design and Engineering Applications} \\ 
\medskip {\LARGE Project Long-term Plans}
\end{center}
\bigskip

\noindent {\bf \large 1. Longer Term Scientific Growth:}

The scientific challenges addressed by this proposal are closely 
related to two recent important developments at McMaster 
University and the University of Waterloo. The first development 
is the establishment of the {\bf McMaster School of Computational 
Engineering and Science (MSCES)}. This school, spearheaded by 
Founding Director Professor Tamas Terlaky, is aimed at creating a 
multidisciplinary environment to promote collaborative research 
and train university graduates to effectively apply the growing 
computing power of technology to the numerous problems arising 
from applications in engineering and science. The urgency of this 
need has recently been highlighted in the United Stated through 
the NSF Atkins report, and the Presidential report on the 
importance of computational science. One of the three major 
research fields of the MSCES, Computational Optimization: Design 
and Control, is directly relevant to the scientific content of 
this project. The recent developments in computational 
optimization have highlighted the importance of effectively 
integrating high performance computational methods with the 
leading-edge research activities in the engineering and physical 
sciences. 

The second development is the establishment of the {\bf 
Electricity Markets Simulation and Optimization Laboratory 
(EMSOL)} at the University of Waterloo.  This initiative, led by 
Professors Miguel Anjos and Kankar Bhattacharya, arises from the 
current trend towards deregulation of the power industry. The 
goal of deregulation is to enhance competition and provide 
customers with a range of choices and greater economic benefits. 
The process has necessitated the reformulation of established 
models for power system operations.  Simultaneously, issues such 
as system reliability, control, and security, are being 
scrutinized and debated in this new deregulated environment. In 
Canada, several provinces have recently started restructuring 
their power sector, and new electricity markets have already been 
established in some of them. There is therefore an urgent need 
for researchers to address issues such as optimal generation 
scheduling, and optimal power pricing in a real-time setting. 

Both of the MSCES and EMSOL initiatives are strongly supported by 
their respective universities.  McMaster University is investing 
significant resources to affirm the status of the MSCES as a 
flagship graduate school. Resources already committed include one 
new faculty position, and a budget of about \$100,000 for the 
first year of operation. Once accreditation is complete, funds 
for student support of up to \$250,000 are also expected for the 
first year of student intake (2006/07).  When a steady-state is 
reached the MSCES is expected to have up to 70 graduate students, 
a corresponding space commitment of over 4,000 sq.ft and up to 
five additional faculty appointments.  

At the University of Waterloo, the EMSOL will be housed in a 
newly renovated space of over 1200 sq.ft which can support the 
activities of 20 graduate students.  The University of Waterloo 
has committed nearly \$50,000 for the renovation of this area. 
The high-performance computing needs of this research have been 
met by a major CFI New Opportunities Grant to the two lead 
researchers, totalling over \$450,000.  Student support is 
currently being provided by various sources, including our 
partners Bell Canada ({\tt http://www.bell.ca}) and ABB ({\tt 
http://www.ABB.us}) from industry, and MITACS and CITO from the 
government sector.

\pagestyle{myheadings} \markboth{T. Terlaky - H.P.O. and App.: 
Long Term Plans}{T.  Terlaky - HPO and App.: Long Term Plans}


The timing of these two initiatives coincides with an increase in 
collaborative work between researchers at these two institutions, 
much of which is already being supported by MITACS.  The support 
by MITACS of this proposal will allow us to bring new members 
from both universities into our team. These new members 
strengthen both the mathematical optimization (Anjos, Deza, 
Fuller) and engineering optimization (Ca\~nizares, Bhattacharya) 
sub-fields. Together with our collaborators from the University 
of Windsor, it is our expectation that the interactions that will 
take place thanks to the support of this proposal will build a 
collaborative research and training network which will be 
sustained through these two major initiatives for many years 
beyond the termination of NCE funding. 

Finally, it is expected that around the time the NCE funding 
terminates, the EMSOL will have established itself as a strong 
research unit, and that it will expand into a self-sustained {\bf 
Center for Electricity Markets and Optimization Research} at 
Waterloo. The resulting synergy with the MSCES and the {\bf 
Advanced Optimization Laboratory (AdvOL)} at McMaster will form a 
world-class research and HQP training cluster in the area of 
Optimization and Engineering Applications. 

\bigskip \noindent {\bf \large 2. Funding Growth Plan:}
   \begin{tightenumerate}
   \item[Industry and Other Partners:]
Our research is currently funded by Rogers and Bell Canada. The 
activities of the EMSOL will bring funding from ABB to support 
this research. Although this support is typically only granted 
for one year at a time, our track record with MITACS so far shows 
that we have always been able to generate sufficient matching 
funds to fully take advantage of our MITACS allocation. 

   \item[Government Partners:]
Beyond the matching of our industry funding through MITACS, we expect to 
generate additional funds from Ontario sources through CITO and/or the new Energy Centre. 

   \item[MITACS:]
As mentioned above, 
our project team is expanding through the addition of new members from both 
McMaster (Deza) and Waterloo (Anjos, Bhattacharya, Ca\~nizares, Fuller). 
This increase in the number of researchers comes with a corresponding increase in 
the number of graduate students and postdoctoral fellows that will be involved 
in the project. 
Much of this additional research is also accompanied by additional industry funds, 
and this is the rationale for our request 
in this proposal for an increase in funding from MITACS. 

Specifically, 
{\bf we are requesting an increase in the base funding to \$220,000 per year, 
plus a yearly allocation of \$30,000 in internship funds specifically to our project.} 

\end{tightenumerate}

\bigskip \noindent {\bf \large 3. New HQP, Networking and 
Technology Transfer Plans:}

We intend to continue a number of networking and training 
activities that have been carried out regularly in recent years. 
These include the weekly student seminar and Advanced 
Optimization Seminar Series at McMaster, the monthly Industrial 
Optimization Seminar Series at the Fields Institute, and the 
Modelling and Optimization: Theory and Algorithms (MOPTA) annual 
conference that moves between the three Universities involved in 
this project (McMaster in 2001 to 2004, Windsor in 2005, Waterloo 
in 2006, McMaster in 2007). 

We also intend to continue organizing special focus research workshops, 
such as the Large Scale Nonlinear and Semidefinite Programming Workshop at Waterloo in 2004, 
the McMaster Optimization Day in 2004, 
the Workshop on Mathematical Programming in Data Mining and Machine Learning 
and the Franco-Canadian Workshop on Combinatorial Algorithms, both held at McMaster in 2005. 
Currently planned special events for the next two years 
include holding a Workshop on Optimization in Engineering at 
the Banff International Research Station in 2006, and hosting the 
triennial International Conference on Continuous Optimization (ICCOPT) at McMaster in 2007. 
The latter will offer HQP training in the form of a Summer School on Computational 
Optimization. 

Apart from all the aforementioned activities, we intend to initiate a new 
series of high-profile guest speakers at the MSCES. 
The launch of the EMSOL will also initiate a series of seminar speakers in 
electricity markets and optimization at Waterloo. 

There has been a marked increase in recent years in 
the collaboration between members of our project. 
For instance, Professor Vannelli spent a sabbatical leave at McMaster in 2005, 
and together with Professor Anjos, they make weekly visits to 
the AdvOL to pursue joint research work. 
Furthermore, 
several of the activities mentioned above are jointly organized by researchers at 
different institutions, and the administration of the project itself is carried out 
across institutional lines. 
For example, although the Project Leader is at McMaster, the web page is hosted at Waterloo. 
Industrial funding has also been obtained from various sources by different project members. 

Finally, we anticipate an expansion of the number of industrial 
internships with our industrial partners, as part of the 
increased opportunities arising from the creation of the MSCES 
and the EMSOL. For example, HydroOne curently supports a Master 
of Engineering Program in Electrical and Computer Engineering at 
Waterloo. The University of Waterloo is now discussing with 
HydroOne broader support of research activities in sustainable 
energy, energy markets, and optimization. This justifies our 
request above for an allocation of MITACS internship funds 
specifically to our project.

\bigskip \noindent {\bf \large 4. Technology Spin-offs:}

The development of software for optimization and engineering will 
continue with respect to McIPM and its variants: a free, open 
source, interior-point optimization package. 

Recent initiatives that will mature under this proposal include:
\begin{tightenumerate}
\item[1.] The development of the solver SeDuMi, whose development 
was recently taken over by the AdvOL. We intend not only to keep 
it as one of the best SDO solvers in the world, but also to 
extend its abilities to include a powerful parallel version. 

\item[2.] The testing of our novel fast approximation algorithms 
for VLSI routing problems, and the development of a package that 
can be used to apply them to real-life, large-scale routing 
problems. 
\end{tightenumerate}

\end{document}