Research Summaries

Back Optimization of Sensor Operation for Search, Surveillance, and Rapid Accurate Decision Making in Maritime, Littoral, and Urban Environments.

Fiscal Year 2019
Division Graduate School of Operational & Information Sciences
Department Operations Research
Investigator(s) Kress, Moshe
Szechtman, Roberto
Sponsor Office of Naval Research (Navy)
Summary We propose to develop models and algorithms to support surveillance, intelligence processing and perimeter protection missions. The three topics we will to work on in the next year are:
1. Learning Models for Surveillance of a Mobile Target: In FY19 we will extend our earlier work along the following lines,
a) Precisely quantify the performance improvement obtained by using sample paths with just one hit or miss in the UCB algorithm.
b) Extend the learning model to include scenarios with more than just one miss (say, a hit, miss, miss, hit). The resulting ML optimization problem will thus incorporate the full Markov transition matrix.
c) Leverage recent concentration of measure bounds for MLEs to develop performance guarantees for the algorithm to be developed in b).
2. Models of Intelligence Operations: We will develop approximate policies, rooted in multi-armed bandit theory, that allow the junior analysts to explore intelligence sources and then exploit the most relevant one.
3. Perimeter Protection Against Strategic Intruders: We consider a dynamic decision process where an agent is confronted by several adversarial entities that operate independently. Each adversarial entity produces malevolent output of certain size, which the agent wishes to intercept. At each time period the agent has to decide which adversary to investigate with the ultimate goal of intercepting as much as possible malevolent output. The adversaries wish to minimize the rate of interception by the agent.
Keywords Surveillance intelligence operations multi-armed bandit perimeter protection
Publications Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal
Data Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal