DScheduler: A Deductive Database for Scheduling Building Construction Tasks

by Sanjai Tiwari, Stanford Univ, Stanford, United States,
Ashish Gupta, Stanford Univ, Stanford, United States,

Document Type: Proceeding Paper

Part of: Computing in Civil and Building Engineering


Computer applications in the construction domain, such as construction scheduling and estimation, require a large amount of data. Reasoning with complex data is facilitated by Artificial intelligence techniques while data storage and retrieval is facilitated by database management systems (DBMS). However, reasoning with large amounts of data needs support from both technologies. Interfacing AI systems with engineering databases is a problem, because the two systems use different data representations, commonly referred to a impedance mismatch. In addition, data-exchange between the two systems is inefficient. Deductive database systems address these problems by providing efficient data storage and reasoning. In this paper, we will describe the design and implementation of a scheduling prototype, DScheduler, using deductive reasoning on top of a construction database. Building design applications typically deal with large sets of objects, such as columns and walls; deductive databases reason with the sets, unlike most expert systems that would reason with every set element by element and thereby lose efficiency. Our approach can be visualized as a tight-coupling of the expert systems and database technologies, and it is applicable to applications in other engineering domains.

Subject Headings: Databases | Construction management | Scheduling | Artificial intelligence and machine learning | Buildings | Computing in civil engineering | Expert systems | Systems management

Services: Buy this book/Buy this article


Return to search