Improving the Solving Efficiency of TOY(FD)
This thesis encourages the use of the CFLP(FD) paradigm for tackling real-life Constraint Satisfaction and Optimization Problems. To do so it focuses on the system TOY(FD). The first part of the research improves the solving performance of TOY(FD) by interfacing the C++ CP(FD) solvers of Gecode and ILOG Solver, and by enhancing the language for a better specification of ad-hoc search strategies. The second part of the research presents two real-life applications of TOY(FD), an Employee Timetabling Problem (ETP) and an empirical analysis of the hardness of the classical Bin Packing Problem. The third part of the research uses the ETP to perform an in-depth modeling and solving comparison among the different CP(FD) paradigms, using the systems Minizinc, ILOG OPL, Gecode, ILOG Solver, SICStus Prolog, SWI-Prolog, PAKCS and TOY(FD).
Post-Doctoral Researcher in UCC Insight
Wednesday, 26 November, 2014 (All day)