日時: 2018年8月23日 (木) 15:00-16:00
開催場所: 筑波大学 システム情報工学研究科 社会工学専攻 E棟 3E210室
講演者: Prof. Hung-Lin Fu(National Chiao Tung University, Taiwan)
題目:
Adaptive Algorithms in Group Testing
概要:
The notion of group search was proposed sometime in World War II. Robert Dorfman published a paper to describe such a novel idea. The purpose of the project is to weed out all syphilitic men called up for induction to the war. Testing an individual for syphilis involves drawing a blood sample and analysing the sample to determine the presence or absence of syphilis. However, testing every soldier individually would have been very costly. "Group Testing" idea was then proposed by a coworker of R. Dorfman, D. Rosenblatt.
Basically there are two types of searching algorithms: Adaptive and non-adaptive. If the knowledge about the tests (positive or negative) can be adapted to decide which set of items is to be tested in next step, then the algorithm is called an adaptive algorithm. On the other hand, if all tests are designed and running simultaneously, the algorithm is non-adaptive.
In this talk, I shall introduce some of the known adaptive algorithms and then present a couple of new results we obtain recently. One of them is for the classical model, called “Quaternary splitting algorithm” and the other one deals with complex model, it is known as a “learning hidden graphs” problem.
------------------------------------------------------------------------------------------------------------------------
日時: 2018年8月23日 (木) 16:00-17:00
開催場所: 筑波大学 システム情報工学研究科 社会工学専攻 E棟 3E210室
講演者: Prof. Yuan-Hsun Lo(Xiamen University, China)
題目:
Shift-invariant Channel Hopping Sequences for Cognitive Radio Networks
概要:
In cognitive radio networks, establishing a communication link between a pair of secondary users (SUs) requires them to rendezvous on a common channel which is not occupied by primary users (PUs). Asynchronous channel hopping (CH) sequences with full degree-of-rendezvous (DoR) enable a pair of SUs to meet on every channel without requiring clock synchronization, and hence minimize rendezvous failures due to the appearance of PU signals. In this talk, we revisit the shift-invariant (SI) sequences, which is used to achieve throughput-invariant and zero-packet-loss throughput 1 for a collision channel without feedback, and generalize them to CH sequences. It shows the SI property of the new CH sequences will enjoy the smallest rendezvous-success-rate (RSR) variance in the cognitive radio networks in various collision-avoidance scenarios.