Lunch time seminars
http://www.dcc.fc.up.pt/%7Ejpp/OPTIMIZATION/Optimization/DCC_seminar_series/DCC_seminar_series.html
In October 2011 we started a new series of seminars at the Computer Science Department.<br/>iWeb 3.0.4http://www.dcc.fc.up.pt/%7Ejpp/OPTIMIZATION/Optimization/DCC_seminar_series/DCC_seminar_series_files/FCUP-DCC.jpgLunch time seminars
http://www.dcc.fc.up.pt/%7Ejpp/OPTIMIZATION/Optimization/DCC_seminar_series/DCC_seminar_series.html
Yuji Shinano, ZIB (Zuse Institute Berlin)
http://www.dcc.fc.up.pt/%7Ejpp/OPTIMIZATION/Optimization/DCC_seminar_series/Entries/2016/5/5_Yuji_Shinano,_ZIB_%28Zuse_Institute_Berlin%29.html
c4c7bf49-700d-4d68-90a9-5807ae9812dbThu, 5 May 2016 16:24:11 +0100In this talk, we introduce the Ubiquity Generator (UG) Framework, which is a software framework to parallelize a branch-and-bound based solver on a variety of parallel computing environments. UG gives us a systematic way to develop a parallel solver that can run on large-scale distributed memory computing environments. A success story will be presented where we solve 14 open MIP instances from MIPLIB2003 and MIPLIB2010 using ParaSCIP, a distributed-memory instantiation of UG using SCIP as MIP solver, on up to 80,000 cores of supercomputers.<br/>Benjamin Müller, Zuse Institute Berlin
http://www.dcc.fc.up.pt/%7Ejpp/OPTIMIZATION/Optimization/DCC_seminar_series/Entries/2016/4/7_Filipe_Alvelos_+_Filipe_Brandao__Column_Generation_for_MIP_2.html
93b31ac4-d818-4641-a6fc-50ddfba65c68Thu, 7 Apr 2016 16:19:53 +0100The constraint integer program framework SCIP (<a href="http://scip.zib.de/">http://scip.zib.de</a>) solves convex and nonconvex mixed-integer nonlinear programs (MINLPs) to global optimality via spatial branch-and-bound over a linear relaxation. Besides being one of the fastest MINLP solvers available in source code, SCIP can also be used as a branch-cut-and-price framework. Furthermore, its plugin-based design allows to extend the framework to solve even more different kinds of problems and to customize the optimization process.<br/><br/>We will present new algorithmic ideas for optimality-based bound tightening and a generalization of the so-called Lagrangian variables bounds. We evaluate the performance impact of these new algorithmic ideas on the MINLPLib2 instance library. Laura Cavalcante
http://www.dcc.fc.up.pt/%7Ejpp/OPTIMIZATION/Optimization/DCC_seminar_series/Entries/2015/7/16_Laura_Cavalcante.html
5ba0d004-90b6-4626-8756-0e1128b89aebThu, 16 Jul 2015 16:46:09 +0100Optimal management of wind power generation in power systems and electricity markets require wind power forecasts, used as inputs to support decision-making problems. In order to improve the forecast accuracy and properly address the stochastic nature of the wind power, a spatio-temporal framework, based on Vector Autoregressive (VAR), can be applied. The forecasting process involves the choice of a suitable sparse model structure for the optimization problem and an optimization algorithm to solve it. In this context, this presentation briefly discusses some paths toward forecast improvement with particular emphasis in the alternating direction method of multipliers (ADMM), a simple but powerful algorithm that is well suited for a wide variety of large-scale distributed modern problems.Mariana Gil (paper by V.Quintana & al.)
http://www.dcc.fc.up.pt/%7Ejpp/OPTIMIZATION/Optimization/DCC_seminar_series/Entries/2015/7/3_Filipe_Brandao_%28paper_by_L.Wolsey%29_2.html
749495ae-fa34-4932-ba6a-258331bafd1bFri, 3 Jul 2015 16:40:38 +0100An introduction go interior point methods for solving nonlinear programming problems in power-engeneering applications.Filipe Brandão (paper by L.Wolsey)
http://www.dcc.fc.up.pt/%7Ejpp/OPTIMIZATION/Optimization/DCC_seminar_series/Entries/2015/6/18_Filipe_Brandao_%28paper_by_L.Wolsey%29.html
b59e36f9-8d0d-4634-9db5-c7df28db483aThu, 18 Jun 2015 16:37:27 +0100Various discrete optimnation problems such as the integer and 0-1 programming problems, and the travelling salesman problem have been represented as discrete dynamic programming, or network problems. We show how such representations lead naturally to a caracterization of the valid inequalities for the feasible solution sets Q of such probles. In particular we obtain polytopes Gamma of valid inequalities having the facets of Q among their extreme points. In addition the problems of "packing" or "covering" with feasible solutions to the discrete problem have natural network representations, which are the duals of problems over Gamma.<br/><br/>Reversing the approach, any special properties of the valid inequalities can in turn be used to give new formulations of the corresponding network problems. In particular this allows a reformulation of the "minimum equivalent knapsack inequality" problem, and the "cutting stock" problem.<br/>