Other current interests include the structure of polyhedra related to resource-constrained scheduling. With colleagues, I have developed specialized cutting planes for so-called `big M' constraint disjunctions, and we are currently working on implementation. I am also interested in the application of randomisation and restart in the context of LP and ILP problems.

My interest in ILP algorithms grew from work in the early 1980's which expressed register-transfer level digital logic synthesis problems using a mixed-integer linear programming formulation. The model contains activities which are optional and have continuously variable duration, and the objective is to minimize both the schedule length and equipment cost. This same characterization encompasses a broad range of industrial processes. IP algorithms in the 1980's were not powerful enough to solve these models. Being young and enthusiastic, I said `How hard can it be?' and began writing an ILP code. BonsaiG was the result. It combines traditional LP-based branch-and-bound algorithms from Operations Research with arc consistency techniques from Artificial Intelligence, exploiting the interplay between techniques developed for continuous optimization and discrete constraint satisfaction. Along the way I discovered polyhedral theory and cutting planes, and the sheer size of the task eventually led me to throw in with the COIN-OR Project.

I also have a lingering interest in state-space search and economic equilibrium models for optimization and constraint satisfaction problems, and more generally in the connections between AI and OR approaches to optimization.