Portfolio · Software
Numerical implementations and visualization workflows supporting my research in dynamic optimization and control.
The code behind this work is being prepared for release alongside the papers it belongs to. If you would like to see any of it sooner, or discuss applying these methods to a problem of your own, please get in touch.
Open GitHub ↗Code that solves the optimality conditions in continuous time, finds the sequence of active constraints and the switching times for a given initial state, assembles the critical regions, and compares the resulting maps with discrete-time ones. Developed for the journal paper.
Tools that test when a critical-region boundary is a hyperplane and construct it analytically. Developed for the CDC paper.
An implementation that reads the sequences of active constraints on the two sides of a boundary and decides, before any computation, whether the boundary is a hyperplane or curved. Developed for the ACC paper.
Code that computes a curved critical-region boundary exactly, with its ends and the regions adjacent to it, without solving an optimal control problem. Developed for the ECC paper.
Tools for classifying parameter points, computing region boundaries exactly, building consistent maps, and visualizing the resulting control structure.
Benchmarks against a discrete-time multiparametric solver on evaporators in series: region counts, build time, memory, the error in the applied input, and constraint violation between discretization instants. Developed for the ESCAPE paper.
Workflows that count and compare the control laws a process controller has to store under the continuous-time and discrete-time formulations. Developed for the FOCAPO paper.
Browser-based experiences and visual explanations that make optimization and control ideas easier to understand without removing the mathematical foundation.