OSIRIS: Optimal Strategy to Innovate and Reduce Energy Consumption in Urban Rail Systems. Project 284868, VII Framework Program. European Commission.

 

This project, in the context of the VII Framework Programme of the European Union, runs from January 2012 until the end of 2014 and involves the CMM and17 European research institutions and companies. The applied research is focused to the study and modeling of energy balances in urban rail transportation systems. The model includes all the variables and  parameters for simulation and optimization of the system, including 4 subsystems: efficient speed curves between two stations, traction and forces, energy balance of the global system, and thermal impacts om passengers. This holistic model has been implemented in a software tool for simulation and optimization as a support for operational decisions.

 

Energy balance for urban and inter-urban rail transportation systems (2013-2014).

This a joint initiative with the Chilean Agency for Energy Efficiency (AChEE), to propose strategies for the reduction of energy consumption in rail transportation. The model includes global the energy balance and permits to identify sources of (in)efficiency. In particular, we are developing new models to estimate the speed curves of trains between two stations. Both cases, urban underground and open air inter-city trains, are analyzed in this project, and applications on real world data form Chilean railway companies are being developed.

 

Models for open pit mining extraction.

The project, funded by FONDECYT, aims to propose for the block sequencing of the extraction from an open pit mine.  We study the discrete and the continuous models, with special emphasis  on the algorithm development and computer implementation. A crucial problem in the mining industry is to determine the optimal sequence of extraction of blocks, in which the mine has been structured for exploitation. Typically a mine can be constituted by several thousands or millions of blocks. The sequencing models for this structure are very complex giving rise to very large combinatorial linear models. Operational mine plans are usually produced on a yearly basis and further scheduling is attempted to provide monthly, weekly and daily schedules. In the first part of this project, an integer programming model is presented to generate pit designs under exposed ore reserve requirements, as an extension of the classical optimization models for mine planning. For this purpose, we introduce a set of new binary variables, representing the extraction, wasting and processing decisions. The model is being coded and tested in a set of standard instances, showing very encouraging results in the generation of operational sequences of extraction and destination of blocks. See More

 

Crew scheduling model in train transportation.

This project was developed from 2008 to 2011. This project generated a model for the crew scheduling problem for train operations, based on a rotated schema, where weekly trips are fixed along the time, meaning that trips program doesn’t vary along the weeks. Each schedule generates a 0-1 medium/large size (hundreds to thousands of variables and constraints) optimization problem, with some complexity on constraints. The special feature of this model is an infinite horizon schedule, due to the rotated schema, where every crew takes the place of the consecutive crew when a new week starts. The problem resolution is performed through three steps: first, finding a feasible solution of infinite length, where schedules repeat between crews; then, an adapted local search strategy is applied to improve the initial solution, in order to equilibrate the weekly working hours among crews and break the symmetry of the problem; finally, drivers are assigned to the scheduled weeks, by solving a flow problem. This strategy has been implemented for real cases applications in High Performance Computing (HPC).

 

Optimization modeling for resource allocation in the Chilean public education system.

The project, developed from 2008 to 2013, presents an economic decision model for planning in public education, considering the satisfaction of the total demand for education in a municipality and the annual cost involved in the system. Given a set of established schools in a municipality and possible new locations, the aim of the model is to find the optimal supply in terms of which schools should be kept open or closed, which grades (or year levels) should be made available and the number of classes that should be provided within each such grade to meet student demand. The model includes specific information about curriculum requirements in terms of the number of hours per year per subject to be provided for each grade level. The objective function essentially seeks to minimize the total fixed and variable costs to schools, which mainly relate to human resources (principal and direction, teaching, administrative and support staff). The integer optimization model for solving this problem was applied to two large municipalities in Chile.