Discrete Event and Hybrid Systems
Infinitesimal Perturbation Analysis (IPA):
- Sensitivity estimation of mean system time of a GI/G/1 queue with respect to arrival and service rates.
Modeling a manufacturing system as a hybrid system:
Coverage control simulator:
- For mission spaces with obstacles
- The use of centralized boosting functions to escape local optima.
- With second-order agents under velocity and acceleration constraints.
- With second-order agents under velocity and acceleration constraints with obstacles in the environment
- The use of decentralized boosting functions to escape local optima with convergence guarantees.
- The use of centralized, decentralized as well as iterative greedy methods to obtain performance guarantees.
- Deciding the optimal agent team composition out of a given set of heterogeneous agents.
Persistent monitoring on graphs:
- Use of IPA to optimize a threshold based routing policy utilized by a set of agents to monitor a set of target points.
Data Analysis and Network Optimization