CSEN 912

4 lecture hours


4  ECTS credits

Multi-Paradigm Optimization (Elective)

Pre-requisites:

Abstract

  • The industrial and commercial worlds are increasingly competitive, requiring companies to be more productive and more responsive to market changes. As a consequence there is a strong need for solutions to large scale combinatorial optimization problems in domains such as production scheduling, transport, resource allocation, finance and network management. Optimization technology is certainly reaching a level of maturity. Having emerged in the 50s within the Operations Research community, it has evolved and comprises new paradigms such as constraint programming and local search techniques. Today, there is a necessity (for efficiency, scalability and tractability reasons) to integrate techniques from the different paradigms.

Outline

  • By the end of this course, the graduate students will have theoretical and practical knowledge of advanced computational algorithms to tackle NP-complete problems from academia and industry. The main objective and goals for this elective if to enrich graduate students with advanced computational techniques and algorithms to tackle complex real world problems arising in industry. It complements the core course on algorithm analysis and design to tackle NP-complete problems.

Goals

  • This elective course covers core methods and techniques from the different paradigms used to tackle constrained optimization problems drawing from the fields of AI and Operations Research. The course addresses the fundamental issue of constrain model design and efficient solution building. This includes:

    • Fundamentals of constraint programming (models, algorithms)
    • Elements of graph theory
    • Linear and integer programming (Simplex, Branch and bound)
    • Local search principles and algorithms
    • Algorithm hybridization

Course Editions

RenewSession