Optimizing Nonlinear Problems through Variable Aggregation
A novel approach to reduce computational complexity in engineering optimization
This research formalizes variable aggregation as a pre-solve technique to create more efficient reduced-space formulations of nonlinear optimization problems.
- Develops a systematic approach for aggregating variables in nonlinear programming
- Introduces an approximate maximum variable aggregation algorithm
- Demonstrates improved computational performance for complex engineering optimization models
- Extends techniques previously applied mainly to linear programming
Enabling more efficient solution of complex nonlinear optimization problems is critical for large-scale engineering applications including process design, power systems, and structural optimization where computational efficiency directly impacts practical implementation.