ARGOS: Dynamic Composition of Web Services for Goods Movement Analysis
and Planning
Project Description
Argos is a flexible data query and analysis system based on the web
services paradigm. As an application domain we will examine several
goods movement planning problems and their effects on spatial urban
structure.
Many scientific problems can be modeled as a workflow that includes
information gathering and processing operations. We propose a unifying
framework where these operations are modeled as web services and the
scientific workflows as compositions of web services. Argos focuses on
automatic composition of web services workflows. Our approach consists
on giving precise logical descriptions to the inputs and outputs of
each web service based on an ontology of the application domain. We do
not assume that the inputs and outputs are simple types, such as
strings or numbers, but that the services process complex types, in
particular, relations. Reasoning with these
relational descriptions allows the Argos planner to
automatically compose a workflow in response to a user data request,
including inserting adaptor services that transform the output of one
service to the input require by another.
We use Argos in a metropolitan transportation planning scenario
in consultation with our advisory team, that includes government
practitioners from the California Department of Transportation, the
Los Angeles County Metropolitan Transportation Authority, the San
Bernardino Associated Governments, the Southern California Association
of Governments, the Port of Long Beach, and others. The scenario
analysis allows evaluation of Argos in terms of its
expressiveness, utility as a transportation planning tool, and ease of
use by developers and practitioners.
Argos in funded by the National Science
Foundation, Digital
Government Research Program .
Participants
Argos is a multidisciplinary project between the
School of Policy,
Planning, and Development (SPPD) and the
Information Sciences Institute (ISI)
of the
University of Southern California (USC)
under the auspices of the
Digital Government Research Center.
-
Prof. Genevieve Giuliano, USC, SPPD
-
Prof. Jose Luis Ambite, USC, ISI
-
Prof. Peter Gordon, USC, SPPD
-
Prof. Qisheng Pan, Texas Southern University
-
Dipsy Kapoor, USC, ISI
-
LanLan Wang, USC, SPPD
-
JiYoung Park, USC, SPPD
Alumni
-
Dr. Hans Chalupsky, USC, ISI
-
Dr. Stefan Decker, USC,ISI, and DERI.
-
Andreas Harth, DERI
- Karanbir Jassar, Yahoo!
-
Mountu Jinwala, Siderean Sotfware
-
-
Abdul Naqeeb Abbasi, Microsoft
-
Matthew Weathers, Biola University
-
Paul Pender, USC, SPPD
Argos Twiki
(restricted access)
Software Demonstrations
Automatic Service Composition with Relational Descriptions
We developed a planner to automatically compose data
processing workflows in response to user's data requests. We use PowerLoom to
represent data sources and operations and to reason about the
relationship between inputs and outputs of services. The planner
automatically inserts adaptor services as necessary. The dg.o2007
paper explains some of the details. Some examples of workflows
automatically generated by the planner:
- Sample workflow
that computes the assignment of freight traffic to highway links in
the Los Angeles Consolidated Statistical Area (LACMSA). The output of
this workflow, as displayed in ArcGis (thicker lines denote higher
truck traffic volume)
- Sample workflow
that computes the intra-regional supply of fuel products (USC commodity
product code 10) for the TAZs in the LACMSA region in millions of US
dollars for the year 2000.
Product Conversion Web Service
A situation that occurs repeatedly when integrating economic data from
different sources is the need to translate data expressed in one
product classification into another. Economic data is reported in a
variety of classifications, including the Standard Classification of
Transported Goods (SCTG), the Standard Industrial Classification
(SIC), the North American Industry Classification System (NAICS),
among many others. Unfortunately, there is no shortage of standard
classifications. Thus, we have developed a Product Conversion Web
Service that automates the translation of data categorized
according to different industry/product classifications. We provide
two ways of accessing the Product Conversion service:
Please send an email to
if you use this web service, and/or you have comments.
The bridge tables used in the web service were developed by JiYoung
Park (jiyoungp@usc.edu) and LanLan Wang (lanlanwa@gmail.com). Please
contact them if you have questions on their estimation methodology.
Automatic Composition of Aggregation Workflows
We built two demos that show our automatic composition of aggregation
workflows as described in this poster and in
the dg.o2005
paper. We used RDF as our
representation language and Triple as our reasoning
engine:
- Demo1
implements the example used in the dg.o2005 paper. It is easier to
visualize and understand.
- Demo2
shows how the BPEL4WS workflow shown below (from the dg.o2004 demo)
can be automatically generated. It uses terms from our ontology. It is
harder to visualize since the graphs grow quite large fairly
quickly.
These demos were presented at the National Conference on Digital
Government Research (dg.o2005).
BPEL4WS workflow
The
demo shows a portion of the Argos workflow that estimates the
commodity imports and exports for the Southern California Association
of Governments (SCAG) area
based on Waterborne Commerce of the United States (WCUS) data and Port
of Long Beach statistics. The demo shows a BPEL4WS workflow where
the data sources and the processing steps are implemented as web
services. It was presented at the National Conference on Digital
Government Research (dg.o2004).
Launch the dg.o2004
demo. Click on the ovals to see intermediate data produced at
different steps of the workflow. You can also see the data in its
native RDF representation.
The demo is described in more detail in this poster and the paper
Argos: An Ontology and Web Service Composition Infrastructure for
Goods Movement Analysis below.
[Note: The links above point to a cached version of the dg.o2004 demo.]
Publications
-
Automatically Composing Data Workflows with Relational Descriptions and Shim Services
José Luis Ambite and Dipsy Kapoor.
Proceedings of the 6th International Semantic Web Conference (ISWC-2007), Busan, Korea, November 2007.
-
Automatic Generation of Data Processing Workflows for Transportation Modeling
José Luis Ambite and Dipsy Kapoor.
Proceedings of the 8th Annual National Conference on Digital
Government Research (dg.o2007), Philadelphia, PA, May 2007.
Best Paper Award
-
Argos: A Framework for Automatically Generating Data Processing Workflows (System Demonstration)
José Luis Ambite and Dipsy Kapoor.
Proceedings of the 8th Annual National Conference on Digital
Government Research (dg.o2007), Philadelphia, PA, May 2007.
-
ARGOS: Dynamic Composition of Web Services for Goods Movement Analysis and Planning (Project Highlights 2007)
José Luis Ambite, Genevieve Giuliano and Peter Gordon.
Proceedings of the 8th Annual National Conference on Digital
Government Research (dg.o2007), Philadelphia, PA, May 2007.
-
Data Processing Workflows in the Social Sciences: Representation and Automatic Generation (Abstract)
José Luis Ambite, Dipsy Kapoor, and Mountu Jinwala.
Proceedings of the 7th Annual National Conference on Digital
Government Research (dg.o2006), San Diego, California, May 2006.
-
Automatic Composition of Aggregation Workflows for Transportation Modeling (pdf)
Jose Luis Ambite and Matthew Weathers.
Proceedings of the 6th Annual National Conference on Digital
Government Research (dg.o2005), Atlanta, Georgia, May 2005.
-
Argos: Dynamic Composition of Web Services for Goods Movement Analysis and Planning (pdf)
Jose Luis Ambite and Genevieve Giuliano and Peter Gordon and
Andreas Harth and Karanbir Jassar and Qisheng Pan and LanLan Wang
Project Highlight at the National Conference on Digital
Government Research (dg.o 2004),
Seattle, Washington,
May 2004.
-
Argos: An Ontology and Web Service Composition Infrastructure for Goods Movement Analysis (pdf)
Jose Luis Ambite and Genevieve Giuliano and Peter Gordon and
Andreas Harth and Karanbir Jassar and Qisheng Pan and LanLan Wang
System Demosntration at the National Conference on Digital
Government Research (dg.o 2004),
Seattle, Washington, May 2004.
-
Integrating Heterogeneous Sources for Better Freight Flow Analysis and Planning (pdf)
Jose Luis Ambite, Genevieve Giuliano, Peter Gordon, Qisheng Pan,
and Sandipan Bharracharjee
Second National Conference on Digital Government Research
(dg.o 2002), Redondo Beach, California, May 2002.
A longer
version of this paper was presented at the Transportation
Research Board 82nd Annual Meeting, Washington, D.C., January 2003.
This paper presents our initial previous work in defining a computational
workflow for the Gordon and Pan commodity flow estimation model.
Presentations