Exploring trade-offs between competing requirements to discover new possibilities in engineering design.
Learning from trade-offs
Designing a rocket or spacecraft involves many requirements: efficiency, reliability, mass and cost, among others. Aerodynamics, structures, propulsion and control are connected, and improving one objective can affect another.
We combine multiobjective evolutionary computation with numerical simulation to investigate these relationships. Rather than finding only one answer, we explore sets of design alternatives and seek to understand what distinguishes them.
Developing methods and applying them to engineering
- Multiobjective evolutionary algorithms and constraint handling
- Topology optimization for shape and structural exploration
- Archiving nondominated solutions and extracting design knowledge
- Applications to aerospace engineering design problems
Past applications include aerospike nozzles, spacecraft trajectories and air-assisted sounding rockets. Published benchmark problems and datasets are available through our resources and archives page.
