City planning and cluster computing


“If you really want to make a friend, go to someone’s house and eat with him… the people who give you their food give you their heart.” – Cesar Chavez

Could better city planning actually improve community cohesiveness, economic output, and quality of life? That’s the belief of Steven Farber, a geography professor at the University of Utah whose research focuses on identifying the characteristics of a region that contribute to the frequency and quality of social interactions. With computer modeling, Farber and his colleagues modeled 80 hypothetical areas, each with 1 million people living and working in them. Each area had a unique allocation of residential and work sites and commuting flow. The team identified all possible scenarios of movement from work to home and determined the potential for a person to meet up with another person after work. This research could help in the planning and construction of urban environments that maximize opportunities for interaction and thereby foster greater social benefits encourage community cohesiveness.

Professor Farber worked with ACI-REF Wim Cardoen to adapt his Matlab code to an MPI/C++ code to make use of CHPC cluster resources. This allowed him to do computations which were taking months on a desktop to be completed in minutes.

Image by Tijs Neutens, Steven Farber, Matthias Delafontaine, Kobe Bossauw in the journal Computers, Environment and Urban Systems 41 (2013) 318-331.
Image by Tijs Neutens, Steven Farber, Matthias Delafontaine, Kobe Bossauw
in the journal Computers, Environment and Urban Systems 41 (2013) 318-331.
Image by Tijs Neutens, Steven Farber, Matthias Delafontaine, Kobe Bossauw
Image by Tijs Neutens, Steven Farber, Matthias Delafontaine, Kobe Bossauw