Faculty of Graduate Studies Online Theses
http://hdl.handle.net/10222/10559
Sun, 19 Nov 2017 12:52:21 GMT2017-11-19T12:52:21ZUnderwater Communication Through Magnetic Induction(MI)
http://hdl.handle.net/10222/73429
Underwater Communication Through Magnetic Induction(MI)
Sana, Ramadan
Wireless Underwater Communication Networks (WUCNs) have recently become a hot topic of research due to applications such as mine detection, navigation, and pollution monitoring. One-third of the earthâ€™s surface is covered by water, making it a challenging environment for communications. Acoustic waves are currently the most common technology used in underwater communications, but this approach suffers from large attenuation and propagation delays. Less commonly used are electromagnetic waves (EM), which experience range limitations in water, and optical waves, which encounter scattering. In this thesis, we will focus on magnetic induction-based communication. Magnetic induction (MI) has several advantages compared to the commonly used acoustic, EM, and optical communication methods. For instance, MI does not suffer from multi-path fading or scattering, and the signal propagation delay is negligible. We will roughly estimate the achievable range, operation frequencies, bandwidth, path loss, capacity and distortions of the MI in conductive media such as water.
http://hdl.handle.net/10222/73429An Experimental Study of Dynamic Biased Skip Lists
http://hdl.handle.net/10222/73416
An Experimental Study of Dynamic Biased Skip Lists
Zhao, Yiqun
Skip lists allow search, insert, and delete operations in O(log n) average time. However, they show a lack of efficiency when dealing with data exhibiting locality of reference, meaning the keys are searched or updated non-uniformly. The dynamic biased skip list (DBSL) allows O(log r(k)) amortized time per operation, where r(k) is the number of distinct keys accessed after k was last accessed.
The performance of DBSL has not been fully evaluated. We implemented DBSL and compared it with move-to-front lists, red-black trees and splay trees, using real-world and synthetic data. Our results show that when a highly biased data sequence contains delete operations, DBSL is more efficient compared to the others. When the degree of bias is moderate, DBSL is not competitive against splay trees, but is more efficient than red-black trees and skip lists. When the data sequence is less biased, DBSL is the least efficient.
http://hdl.handle.net/10222/73416EXPERIMENTAL INVESTIGATION ON THE OUT-OF-PLANE BEHAVIOUR OF CONCRETE MASONRY INFILLED FRAMES
http://hdl.handle.net/10222/73415
EXPERIMENTAL INVESTIGATION ON THE OUT-OF-PLANE BEHAVIOUR OF CONCRETE MASONRY INFILLED FRAMES
Wang, Chongyang
This thesis presents an integral part of an on-going research on out-of-plane behaviour of masonry infills, this study was motivated to augment the existing experimental database on out-of-plane behaviour and strength of masonry infills by expanding on the variations on the parameters, and to assess the efficacy of existing analytical methods for infill out-of-plane strength calculation. A total of five scaled masonry infilled frame specimens including four masonry infilled RC frames and one masonry infilled steel frame were tested to failure under uniformly distributed out-of-plane loading. Parameters considered for the RC frame specimens included infill opening (door opening), prior in-plane damage, and gaps at either frame-to-beam or frame-to-column interfaces. Experimental results showed that all studied parameters have a significant effect on the out-of-plane behaviour and strength of infills.
http://hdl.handle.net/10222/73415MAXIMAL AND CONVEX LAYERS OF RANDOM POINT SETS
http://hdl.handle.net/10222/73405
MAXIMAL AND CONVEX LAYERS OF RANDOM POINT SETS
Cuong Nguyen
This thesis studies two well-known geometric structures in computational geometry: maximal points and convex hull. Extending the concepts to multiple maximal and convex layers is natural. We study the maximal and convex layers of a point set in $d$ dimensions drawn from a uniform or component-independent (CI) distribution. A distribution is component-independent if each coordinate of a point is chosen independently from a continuous distribution. Precisely, we want to compute and to bound the expected size of the first $k$ layers. For the first set of results, we show that, for $d \in \{2,3\}$, the first $n^{1/d-\epsilon}$ maximal layers can be computed using $dn + o(n)$ scalar comparisons with high probability. For $d \ge 4$, the first $n^{1/2d-\epsilon}$ maximal layers can be computed within this bound with high probability. The first $n^{1/d-\epsilon}$ convex layers in two dimensions, the first $n^{1/2d-\epsilon}$ convex layers in 3D, and the first $n^{1/(d^2+2)}$ convex layers in $d \ge 4$ dimensions can be computed using $2dn + o(n)$ scalar comparisons with high probability. Since the expected number of maximal layers in 2D is $2\sqrt{n}$, our result for 2D maximal layers shows that it takes $dn + o(n)$ scalar comparisons to compute a $1/n^\epsilon$-fraction of all layers in the average case. For the second set of results, we show that the $k$th maximal and convex layer of a point set drawn from a continuous CI distribution in $d$ dimensions has expected size $O(k^d \log^{d-1} (n/k^d))$.
http://hdl.handle.net/10222/73405