Contact me by email if interested.

GPU and Multi-Core Computing

Todays optimization software does not utilize the latest multi-core and GPU hardware well. Time consuming optimizations often run on a single processor. It is a major challenge to unleash the power of high-end hardware to solve large instances of combinatorial optimization problems, such as routing, scheduling, packing and planning etc. Furthermore, massive amounts of data are becoming available online, from social networks and elsewhere, and making use of relevant sources in realtime for planning and optimization is another major challenge.

General Research Interests

• Evolutionary Computation,
• Heuristic Methods,
• Multi-Objective optimization,
• Vehicle Routing,
• Cutting and Packing,
• Distribution network optimization,
• Combinatorial Optimization (in general),
• Parallel algorithms - GPU, multi-core and HPC,
• Fast/Big Data,
• Smart Cities.

SampleProjects

• Modelling user behaviour in public transport systems,
• Integrating multiple data sources for transport planning,
• Workforce Scheduling,
• Lot sizing problems how much product to manufacture in one go, or how much inventory to hold,
• Using GPUs to speed up algorithms for combinatorial optimization,
• Mapping multi-objective algorithms onto multi-core hardware.

Recent PhD students have undertaken projects in the following areas:

• Bus Routing and Scheduling,
• Speeding up combinatorial optimization using GPUs,
• Evolutionary algorithms for vehicle routing,
• Facility location/allocation for distribution networks,
• Metaheuristics for vehicle routing problems with time-dependent travel time.

Back to home page