Invitation to Submit Scholarly Articles Using Agent-Based Simulation to Tackle Challenging Civil Engineering Problems

Abstract

I was honored to be invited recently by editors Jim Garrett and Lucio Soibelman to join the Journal of Computing in Civil Engineering Editorial Board as an associate editor. My role on the Editorial Board will be twofold: 1 to provide editorial support for submitted manuscripts that utilize agent-based simulation and multiagent simulation methods to address complex civil engineering problems and 2 to encourage more submissions by civil engineering faculty employing these methods. A small but growing community of scholars is using agent-based simulation methods to tackle a range of difficult civil engineering problems, including building occupant user behavior Dijkstra and Timmermans 2002 , claims negotiation Ren et al. 2002 , construction management pedagogy Rojas and Mukherjee 2006 , construction operations Mohamed and AbouRizk 2005 , cross-cultural differences in global projects Horii et al. 2005 , dispute resolution El-adaway and Kandil 2010 , evacuation of buildings Shi et al. 2009 , project learning Taylor et al. 2009 , project organizations Jin and Levitt 1996 , project scheduling Christodoulou 2010 , structural design Soibelman and Pena-Mora 2000 , supplier sourcing Ng and Li 2006 , and traffic and equipment flow Kim and Kim 2010 . Axelrod 1997 describes the value of simulation as being rooted principally in prediction, proof, and discovery. But he notes that prediction is the form of simulation most people think of when they think about simulation in the research context. The use of simulation for discovery is “at least as important as proof or prediction” Axelrod 1997 . Agent-based simulation is often used to discover important relationships or propositional rules from patterns in the simulated output from relatively simple models. In agent-based simulation the complexity should be in the simulated results, not the model assumptions. If, for example, a simulation is developed to predict the schedule duration of a project, then the assumptions may need to be complex to ensure an accurate prediction. If, on the other hand, the purpose of the simulation model is to understand the fundamental organizational interactions or dependencies that lead to schedule growth, then the assumptions should be simple and focus narrowly on the important elements of the interaction. Including too many details about the actual interaction may detract from the researcher’s ability to identify patterns that bear on the research question posed. It is also important to note a distinction when comparing predictive simulation with discovery-driven agent-based simulation. In general, simulation validity depends on the purpose of the model Lehman 1977 . When agent-based simulation methods are used for discovery, validation strategies should focus on the validity of the elements taken from other models that are used to formulate the model internal validity , the validity of any