Aeon: Synthesizing Scheduling Algorithms from High-Level Models

Monette, Jean-Noël;Deville, Yves;Van Hentenryck, Pascal

Files

Aeon.pdf
  • Open Access
  • Adobe PDF
  • 244.31 KB

Details

Authors
  • Monette, Jean-NoëlUCLouvain
    Author
  • Deville, Yvesorcid-logoUCLouvain
    Author
  • Van Hentenryck, PascalBrown University, USA
    Author
Abstract
This paper describes the AEON system whose aim is to synthesize scheduling algorithms from high-level models. AEON, which is entirely written in COMET, receives as input a high-level model for a scheduling application which is then analyzed to generate a dedicated scheduling algorithm exploiting the structure of the model. AEON provides a variety of synthesizers for generating complete or heuristic algorithms. Moreover, synthesizers are compositional, making it possible to generate complex hybrid algorithms naturally. Preliminary experimental results indicate that this approach may be competitive with state-of-the-art search algorithms
Affiliations

Citations