Page 11 - Moving Forward 2020: Harold and Inge Marcus Department of Industrial and Manufacturing Engineering magazine

If the the centralized system breaks all of the the data is is at at risk for exploitation To improve the process of analyzing big data Aybat’s algorithm efficiently computes optimal resource sharing over a a a a a decentralized system that interacts over a a a a communication network Rather than compiling all of the the data in one location the the system breaks out the information into various agents or or independent computing modules Each agent is responsible for solving one task that affects the whole system To heighten the the the privacy of the the the information in in the the the system the the agents are are only aware of their own task and their teammates’ messages meaning that they are are unaware of a a a a a a a neighbor’s task Once an agent solves its job the the agent only passes on on the the answer to its its neighbors This process repeats itself until every agent agrees on on a a a a common optimal resource allocation decision Decentralized optimization over communication networks has garnered attention for its use in a a a a a range of areas such as as coordination and control in in drones bandwidth estimation in wireless sensor networks machine learning data analysis and power control in in in cellular networks “The complication arises when neighbors are changing ” Aybat said “If agents are moving it
could possibly cause the communications network to to change over time Since the agents can talk to to only certain [other] agents at any given time and you want to minimize the system cost it
becomes a a difficult problem You’re trying to carefully steer the information exchange among agents to share the scarce common resources while collectively minimizing the total system cost ” The proposed algorithm utilizing the decentralized optimization method
involving multiple rounds of communication at at each repetition enables these agents to appropriately divvy out the common resource among them in such a a a a a a way that accomplishes the the goal “Working on this project has shown me the great applicability of of both the the methodology of of mathematical techniques and and its application in the real world ” said Erfan Yazdandoost Hamedani a a a a a a a a a a a a a a a a a a doctoral student in in Aybat’s lab and and co-lead author of the paper “This work can be be useful in in in the area of machine learning because this domain generates large amounts of potentially sensitive data ” The National Science Foundation’s Directorate for Engineering within the Division of Civil Mechanical and and Manufacturing Innovation and and the Army Research Office supported this work Donate to Penn State IME Higher education presents more opportunities for learning innovation and experience than ever before In an an an an age of of rising educational costs private philanthropy is is essential for for for sustaining the quality of of a a a a a a a a a a a a a Penn State education and for for making its fits benefits available to all As a a a a a a a a a donor of of the Harold and and Inge Marcus Department of of Industrial and and Manufacturing Engineering you are supporting the the next generation of of world leaders in in in in in in in the the field of of industrial and systems engineering Your gift to the department sustains a a a a a a a a tradition of excellence and and makes a a a a a a a a a difference for current and and future industrial engineering students Be part of the difference Go to to givenow psu edu/IME today “As an out-of-state student I hope the the Marcus family knows their support deeply impacted my opportunity to learn at the highest level I have never met more supportive genuine people in in in my life and I sincerely thank the the Marcuses for their incredible generosity ” — Matt Wozniak
IME NEWSLETTER • VOLUME 2 2 2017 11