365
International Journal on Advances in Intelligent Systems, vol 10 no 3 & 4, year 2017, http://www.iariajournals.org/intelligent_systems/
2017, © Copyright by authors, Published under agreement with IARIA - www.iaria.org
Semantic Behavior Modeling and Event-Driven Reasoning for Urban System of Systems
Maria Coelho and Mark A. Austin
Department of Civil and Environmental Engineering,
University of Maryland, College Park, MD 20742, USA
E-mail: memc30@hotmail.com; austin@isr.umd.edu
Mark Blackburn
Stevens Institute of Technology,
Hoboken, NJ 07030, USA
E-mail: mblackbu@stevens.edu
Abstract—Modern urban infrastructure systems are defined by
spatially distributed network structures, concurrent subsystem-
level behaviors, distributed control and decision making, and
interdependencies among subsystems that are not always well
understood. The study of the interdependencies within urban
infrastructures is a growing field of research as the importance
of potential failure propagation among infrastructures may lead
to cascades affecting multiple urban networks. There is a strong
need for methods that can describe the evolutionary nature
of “system-of-systems” (SoS) as a whole. This paper presents
a model of system-level interactions that simulates distributed
system behaviors through the use of ontologies, rules checking,
message passing mechanisms, and mediators. We take initial steps
toward the behavior modeling of large-scale urban networks as
collections of networks that interact via many-to-many association
relationships. The prototype application is a collection of families
interacting with a collection of school systems. We conclude
with ideas for scaling up the simulations with Natural Language
Processing.
Keywords-Systems Engineering; Ontologies; Behavior Model-
ing; Mediator; Network Communication.
I. I NTRODUCTION
This paper is concerned with the development of modeling
abstractions, procedures, and prototype software for the be-
havior modeling of urban systems of systems with ontologies,
rules and message passing mechanisms. It builds upon our pre-
vious work [1], [2] on distributed systems behavior modeling
with semantic web technologies.
A. Problem Statement
The past century has been marked by outstanding advances
in technology (e.g., the Internet, smart mobile devices, cloud
computing) and the development of urban systems (e.g., trans-
portation, electric power, waste-water facilities and water sup-
ply networks, among others) whose individual resources and
capabilities are pooled together to create new, more complex
systems that offer superior levels of performance, extended
functionality and good economics. While end-users applaud
the benefits that these systems of systems afford, model-based
systems engineers are faced with a multitude of new design
challenges that can be traced to the presence of heterogeneous
content (multiple disciplines), network structures that are spa-
tial, multi-layer, interwoven and dynamic, and behaviors that
are distributed and concurrent.
Large-scale urban systems do not follow a standard cradle-
to-grave lifecycle. Instead, the constituent domains within a
city evolve over extended periods of time in response to
external forces (e.g., the need for economic expansion) and
disruptive events (e.g., the need for planning of relief actions
in response to a natural disaster). In both cases, planning of
urban operations is complicated by the large scale of modern
cities, the large number of constituent behaviors, and multiple
dimensions of interdependency among physical, cyber and
geographic systems [3]. These facts are what makes cities
“system of systems,” rather than just systems, and they change
the very nature of systems design and management. For exam-
ple, in order for the communication among the participating
urban domains to occur in an orderly and predictable way,
designers need to pay attention to the boundaries (or interfaces)
of domains [4]. Similar concerns exit for the replacement of
aging infrastructure. In his article on the topic of complex
system failure “How Complex Systems Fail,” Cook discusses
how complex systems are prone to catastrophic failure, due to
the impractical cost of keeping all possible points of failure
fully protected, and even identifying them all [5]. When part
of a system fails, or perhaps an unexpected combination of
localized failures occurs, there exists a possibility that the
failure will cascade across interdisciplinary boundaries to other
correlative infrastructures, and sometimes even back to the
originated source, thus making highly connected systems more
fragile to various kinds of disturbances than their independent
counterparts. Figure 1 presents an overview of some generic
interdependencies among key infrastructure sectors: oil and
natural gas, electricity, transportation, water, and communica-
tions.
B. Scope and Objectives
In order to understand how cascading failures might be best
managed, it is necessary to have the ability to model events
and the exchange of data/information at the interdependency
boundaries, and to model their consequent effect within a
subsystems boundary. This points to a strong need for new
capability in modeling and simulation of urban infrastructure
systems as system-of-systems, and the explicit capture of
infrastructure interdependencies. We envision such a system
having an architecture along the lines shown in Figure 2, and
eventually, tools such as OptaPlanner [7] providing strategies
for real-time control of behaviors, assessment of domain
resilience and planning of recover actions in response to severe
events. This paper presents a model of distributed system-
level behaviors based upon the combined use of ontologies,
rules checking, and message passing mechanisms, and explores