Authors:
Alexander Brodsky
;
Shane G. Halder
and
Juan Luo
Affiliation:
George Mason University, United States
Keyword(s):
DG-Query, XQuery, Mathematical Programming.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence and Decision Support Systems
;
Databases and Information Systems Integration
;
Enterprise Information Systems
;
Operational Research
;
Problem Solving
;
Query Languages and Query Processing
;
Strategic Decision Support Systems
Abstract:
Decision optimization is broadly used for making business decisions such as those for finding the best production planning in manufacturing. An optimization model may indicate the total cost of a certain supply chain given the various sourcing and transportation options used; the corresponding optimization problem can be to select among all possible sourcing and transportation options to minimize the total cost. Optimization modelling requires considerable mathematical expertise and effort to generate effective models. Additionally, the optimization process is heavily dependent on data. However, optimization languages such as IBM’s ILOG CPLEX OPL and Bell Laboratories’ AMPL, do not provide native support for manipulation of XML data. On the other hand, XQuery is a language for querying and manipulating XML data, which has become a ubiquitous standard (W3C) for data exchange between organizations; although, XQuery has no decision optimization functionality. To resolve this gap, this p
aper proposes DG-Query, an XQuery-based Analytics Language that seamlessly merges the XML data transformation and decision optimization capabilities. This is accomplished by first annotating existing XQuery expressions to precisely express the optimization semantics, and second to translate the annotated queries into an equivalent mathematical programming (MP) formulation that can be solved efficiently using existing optimization solvers. This paper presents DG-Query with an example, provides its formal semantics, and describes implementation through a reduction to MP formulation.
(More)