This dissertation presents an experimental study aimed at assessing the feasibility of parallelizing
the constraint solving process using Graphical Processing Units (GPU s). GPUs support a form of data parallelism that appears to be suitable to the type of processing required to cycle through constraints and domain values during consistency checking and propagation. The dissertation also
illustrates an implementation of a constraint solver capable of hybrid propagations (i.e., alternating
CPU and GPU) and parallel search, and demonstrates the potential for competitiveness against
sequential implementations. We consider the Protein Structure Prediction problem as a hard
combinatorial real-world problem as case study to show the advantages of combining parallel search
and parallel constraint propagation on a GPU architecture. We present the formalization and implementation of a novel class of constraints to support a variety of different structural analysis of proteins, such as loop modeling and structure prediction.. We demonstrate the suitability of a GPU approach to implement such MAS infrastructure, with significant performance improvements over the sequential implementation and other methods.
