Week -2 to 1
- Familiarized myself with the Port-of-Entry (POE) problem by reading a myriad of papers on the subject. For a complete list of my references and other resources used in this project, please see the References & Links tab to the right.
- Met the POE research group for the first time and began thinking about the current problems to be solved.
Week 2
- Met with the NYC Nuclear Detection research group for the first time and began thinking about the problem of metropolis nuclear detection.
- Began working with Tsvetan Asamov on optimal, randomized layered inspection in the POE detection problem.
Week 3
- Continued working with Tsvetan. In particular, I helped him work out the correct formulas for mapping the previous model's output into the new space optimized by his model for comparison.
Week 4
- Met with Jolie Cizewski, a physicist at Rutgers University, to discuss the physics of radiation processes and detection.
- Created a preliminary shielding model and expressions for the probability of Type II error.
- Presented the preliminary shielding model to the POE research group and listened to critiques and suggestions.
Week 5
- Had a small breakthrough in creating a practical shielding model: By utilizing energy windowing, information about the sensor technology, and, chiefly, the memoryless property of the Poisson distribution I was able to express the actual number of radiation particles observed by the detector in a convenient form.
- Formulated the complete shielding model, and presented this to both the POE group and the NYC Nuclear Detection group.
Week 6
- Began coding the shielding model in Maple. Generated some preliminary results.
- Presented the preliminary results to Professor Elsayed and internalized his critiques and suggestions.
Week 7
- Finished coding the shielding model in Matlab (switched from Maple to gain additional numerical precision). Generated some numerical results.
- Presented my work at the Nuclear Detection research group meeting, the POE research group meeting, and again for the DIMACS REU.
Week 8
- Added optimization techniques to the shielding model program to choose optimal threshold levels to minimize either expected prevented damage, total cost, etc. subject to budget constraints.
- Generated additional numerical results that included optimized threshold levels.
- Wrote up a technical report on my work this summer.