Schedule

The conference is now over!

Soon, there will be some photos available...

CIO Springer TAP Compal BES FCT
Schedule
TB1 Stochastic Optimization I
[CONTRIBUTED SESSION]

Chair: Udatta Palekar, University of Illinois, United States
Room 3.1.7

 

TB1.1
ETEM-SG – A Computational Management Science Model to Assess Energy Transition at Regional Level

Christopher Andrey (christopher.andrey@ordecsys.com), Frederic Babonneau (fbabonneau@ordecsys.com), Alain Haurie (ahaurie@ordecsys.com)
In this paper, we present a long-term capacity expansion model of a regional energy sector at the regional level, which is designed to assess the transition to renewable generation and the potential offered by smart grid operations. The recent development of smart grid technologies considerably enhances the flexibility of the energy systems by allowing shifting some of the loads, thereby adapting the demand to external signals such as the energy price. This mechanism, known as demand-response, has been modelled in ETEM-SG, to represent the flexibility found in water heaters, residential appliances, industrial processes, etc. Moreover, the transition to renewable generation, with the intermittent production from wind and PV cells, imposes the concomitant development of temporary storage capacities such as batteries, ultra capacitors or flywheels. The model considers also the possibility of using electric vehicles’ batteries as temporary storage units. In order to deal with the uncertainties pertaining to different aspects of the energy system, ETEM uses both stochastic programming (e.g. to understand the influence of weather variability on the deployment of renewables) and recent advanced developments in robust optimisation techniques (e.g. to deal with uncertainties in the adoption of electric vehicles). The model has been implemented for the Leman area region in Switzerland.

Keywords: Energy system, Smart Grids, Stochastic and robust optimization

 

TB1.2
Empirical analysis of real instances of a pig supply chain problem by a two-stage stochastic program

Esteve Nadal (enr1@alumnes.udl.cat), Victor M. Albornoz (victor.albornoz@usm.cl), Lluis Miquel Pla Aragones (lmpla@matematica.udl.cat)
This paper presents the formulation of a two-stage stochastic model with the aim to optimize the pig production system according to the new supply chain structure arising in the latest years. The use of the model is illustrated with a case study based on a real instance. The model maximizes the total benefit calculated from the incomes of the animals and the production costs over the time horizon considered. Also, it provides a schedule of transfers between farms, occupancy of facilities and trucks involved. The model uses integer variables, but it is far to find an optimal solution because of the time spent. Taking this into consideration, further analysis has been done by relaxing the integrity of the variables and studying the model behavior when parameters affecting directly to the execution’s time are modified. As a conclusion, recommendations and future ways of research are provided.

Keywords: supply chain, stochastic optimization, pig production process

 

TB1.3
Solving a Stochastic Aircraft Allocation Problem using Parallel Programming

Udatta Palekar (palekar@illinois.edu), Akhil Langer (alanger@illinois.edu), Laxmikant Kale (kale@illinois.edu)
We consider an aircraft allocation problem in the context of military airlift planning under uncertain demand. The problem is modeled as a 2-stage stochastic integer program. We develop a parallel branch and bound algorithm for the problem using the Charm++ programming environment. At each vertex of the branch-and-bound tree we solve a stochastic linear program. We describe the load balancing strategy used to effectively prioritize the computations associated with the various linear programs that must be solved in such an implementation. After discussing preliminary computational results from our implementation, we describe a scenario decomposition technique that allows us to achieve further computational speed-up.

Keywords: Resource scheduling, parallel programming, stochastic programming