Multi-objective
optimisation.
Real engineering decisions are never single-objective. Performance fights weight, reliability fights cost, accuracy fights latency. I build frameworks — like the Multi-Objective Sensor Optimisation Framework (MOSOF) — that map the full trade-off surface and let stakeholders pick from the Pareto front instead of arguing about it.
The front is the answer.
A 3-objective minimisation problem: cost, weight, and risk. Drag to rotate. The orange surface is the Pareto front — every point on it is non-dominated. Black points are dominated by at least one orange point.
MOSOF + NDCI: diagnostic value, quantified.
The Multi-Objective Sensor Optimisation Framework treats sensor selection as a constrained search over a network's information surface. The Normalised Diagnostic Contribution Index (NDCI) — introduced in our 2025 Sensors paper — gives every sensor a comparable score for its share of system-level diagnostic coverage, fixing the long-standing ambiguity of "more sensors = better".
Validated on the Boeing 737-800 Environmental Control System through Cranfield's SESAC platform, the framework consistently identifies smaller, lighter, cheaper sensor suites that match or exceed the diagnostic capability of larger reference configurations.
- 01Define objectives. Cost, weight, reliability, information, latency — whatever the stakeholders fight about.
- 02Encode constraints. Physical, regulatory, supply, schedule.
- 03Search. Genetic algorithms over the candidate space; NSGA-II family by default.
- 04Front + knee. Deliver the Pareto front and a knee-point recommendation with explicit trade-off context.
Performance
Information gain · diagnostic coverage
Weight
Mass · footprint · payload
Reliability
MTBF · failure rate · availability
Cost
Acquisition · install · lifecycle
Bring me a hard trade-off.
If you're stuck choosing between options that each look optimal on one axis and broken on another, that's exactly the problem this method is built for.
hello@buraksuslu.com →