Scalable simulation for health polices


The purpose of this project is to improve a preliminary Individual-based model of tuberculosis spreading, its implementation using parallel computing and its adaptation for the simulation of tuberculosis dynamics in Barcelona.

Tuberculosis is still one of the most important infectious diseases worldwide, killing over 2.5 million people each year. It is estimated that one third of world population has already been infected by Mycobacterium tuberculosis. In particular, the Ciutat Vella neighbourhood in Barcelona has a tuberculosis incidence which is comparable with the incidence in countries like Sudan.
Classically, most of the models that have been built to address tuberculosis epidemiology follow a structured top-down strategy, i.e., they divide the population into different classes (e.g., susceptible, exposed, infected and recovered in the case of a SEIR model) and fix specific fluxes between these groups. Nevertheless, due to the small percentage of sick people among infected (in average, only a 10% of infected would develop an active disease), when the studied population is not big enough the resultant number of sick individuals can question the use of differential equations.
In this case, which would be the case of a city like Barcelona, the most appropriate approach is the individual-based methodology. The biomedical expertise for the model improvement will be provided by the Experimental Tuberculosis Unit from the Institut de Recerca Germans Trias i Pujol – Can Ruti. The technical support for computing issues (simulation framework, parallel computing, high-processing capacity facilities, etc...) will be provided by the inLab. The possibility of running the simulations in the Barcelona Supercomputing Center will be also evaluated.

Objectives to achieve

  • To improve a preliminary individual-based model of tuberculosis spreading in a closed community.
  • To adapt the model for reproducing the Barcelona data of last years.
  • To adapt the model for its implementation in a parallel computing infrastructure
  • To implement the model in a specific framework developed by the InLab, Yades.
  • To use the simulator as a test-platform of tuberculosis epidemiology control strategies for decision-making.


Duration of the project 
December, 2014
- present
MPI (Message Passing Interface),
GNU Octave,
Areas of expertise involved in the project 
Project Manager 
Joan Francesc Gilabert

Follow us on

Els nostres articles del bloc d'inLab FIB


inLab FIB incorporates esCert


inLab is member of