ゼミナール発表

日時: 9月30日(金)3限(13:30-15:00)


会場: L1

司会: 川原 純
笠井 裕貴 1551031: M, 2回目発表 大規模システム管理 笠原 正治,安本 慶一,笹部 昌弘
title : Performance Evaluation of Congestion-Aware Route Selection for Automatic Evacuation Guiding Based on Cooperation between Evacuees and Their Mobile Nodes
abstract : When large-scale disasters occur, evacuees have to evacuate to safe places quickly. For this purpose, an automatic evacuation guiding scheme based on cooperation between evacuees and their mobile nodes has been proposed. The previous work adopts a shortest-distance based route selection and does not consider the impact of traffic congestion caused by evacuation guiding. In this study, we examine the effectiveness of a congestion-aware route selection for the automatic evacuation guiding. We first adopt a traffic congestion model where each evacuee's moving speed on a road is determined by the population density of the road and his/her order among evacuees traveling in the same direction. Based on this congestion model, each mobile node estimates the cost, i.e., traveling time, of each road in the area and selects a route with the smallest cost. Through simulation experiments, we show that the congestion-aware route selection can reduce both average evacuation time and maximum evacuation time compared to the shortest-distance based route selection, especially under highly congested situations. We also introduce future work including application to general traffic networks.
language of the presentation: Japanese
発表タイトル:避難者・モバイル端末連携型自動避難誘導方式における渋滞を考慮した経路選択の性能評価
発表概要:大規模災害発生後,被災者は迅速に安全な避難先へ移動する必要がある.一方,迅速な避難支援を目的に,避難者が所有するモバイル端末を用いた自動避難誘導方式が提案されている.ただし既存研究では,避難所までの最短距離に基づく距離優先型経路選択を用いており,特定の道路への避難者の集中による渋滞が避難時間へ与える影響が考慮されていない.そこで本研究では,自動避難誘導方式における渋滞を考慮した経路選択の有効性を検証する.まず,道路上の人口密度と同方向に移動する避難者内での順番に応じて各避難者の移動速度が決まる渋滞モデルを定義する.この渋滞モデルに基づき各道路の通行時間を示すコストを定義する.そして各避難者は避難所までの総コストが最小となる経路を選択する.シミュレーション評価により,従来の距離優先型経路選択と比較して,渋滞考慮型経路選択は混雑度が増すほど,平均避難時間および最大避難時間の観点で改善効果が高いことを示す.また,今後の研究のために,一般的な交通網への応用を踏まえた今後の展望について述べる.
 
秦 恭史 1551080: M, 2回目発表 ユビキタスコンピューティングシステム 安本 慶一,笠原 正治,荒川 豊,諏訪 博彦,藤本 まなと
title: The discovery of important Twitter accounts by clustering the pre- and post-earthquake social network
abstract: Previous works have shown that Twitter is a useful information distribution tool. To further take advantage of Twitter, finding important accounts with high information diffusion capacities required. Prior works have proposed a method of finding important account by looking at Twitter as a single network. However, the Twitter network includes the connections between similar accounts, and is divided into many clusters. Therefore, if the Twitter network is considered as a single network, it is difficult to find the important accounts that each user needs. In this work, we performed clustering using a fast graph mining technology for large-scale networks, and we extract important accounts using the Personalized PageRank algorithm. Our approach finds important accounts that each user needs faster and more accurately than previous methods.
language of the presentation: Japanese
発表題目: 震災前後のSNSネットワークにおけるクラスタリングに基づく重要アカウントの発見
発表概要: 災害時,情報を広範囲に隅々まで広げることは重要であり,その一つとしてTwitterは情報流通のツールとして有用であることが関連研究により明らかになっている.Twitterをより活かすためには拡散能力の高い重要なアカウントを見つけることが必要である.石原らは,ネットワーク全体を一つのネットワークとして重要アカウントを抽出している.しかしながら,AmacらによればTwitterネットワークは近い人同士でつながりを持つものと考えられ,複数のクラスタに分かれているものと考えられる.そのため,全体で重要なアカウントがすべてのクラスタにおいて重要とは限らない.よって我々は,大規模ネットワークに対応した高速グラフマイニング技術を使用しクラスタリングを行い,パーソナライズドページランクを用いて重要アカウントを抽出する.これにより各ユーザが必要とする重要アカウントを見つけることができ,情報拡散に有効であることを示す.
 
高 尚昊 1551042: M, 2回目発表 インターネット工学 小笠原 司,笠原 正治,安本 慶一,門林 雄基

title:The Improvement of Competition for Bandwidth between MPEG-DASH Clients in the Adaptive Bit-rate Streaming

abstract:HTTP streaming currently dominates the Internet traffic. The standards of HTTP streaming, MPEG-DASH is an adaptive streaming technique introduced to enable a high quality of user experience. However, in the home network, multiple video streams will compete for bandwidth and lead to poor MPEG-DASH performance. This will impact the received quality of user experience. In this study, I research the factors behind this bandwidth competition problem and propose a solution. Within this presentation, I will describe the current progress of my research and the future work.

 
花井 直樹 1551081: M, 2回目発表 インターネット工学 小笠原 司,安本 慶一,岡田 実,門林 雄基
title: Smartphone detection method using multiple device and sensing information
abstract: This study focuses on a smartphone detection method in order to support evacuation activity against flood. To save people from flood, it is significant to support their evacuation adequately. However, to urge them to evacuate, visiting them all is difficult. Then, to support the evacuation activity, this study proposes a smartphone detection method for grasping existence of people in a targeted area. By analysing device (e.g., LTE, Wi-Fi, Bluetooth) information and sensing information, the proposed method detects existence of a smartphone (i.e., people) among various IoT devices.
language of the presentation: Japanese
 

会場: L2

司会: 大和 勇太
VU HOANG GIA 1561036: D, 中間発表 コンピューティング・アーキテクチャ 中島 康彦,井上 美智子,高前田 伸也,TRAN THI HONG

title: *** Checkpointing on FPGA ***

abstract: *** FPGAs provide reconfigurability and high performance for parallel applications. Modern FPGAs can be integrated in computing systems as accelerators so that they can combine with host CPU to execute offload applications. This integration puts more pressure on the fault tolerance of computing systems and the question how to improve the dependability becomes crucial. Similar to CPU-based system, checkpoint/restart techniques are expected to be developed and applied to FPGA-based computing systems. There are two issues rising in this situation: how to checkpoint and restart FPGA, and how this checkpoint/restart model works well with the checkpoint/restart model of the whole computing system. In this work, first we propose two checkpoint/restart architectures along with a checkpointing mechanism on FPGA. Second, we propose “fine-grain” management for checkpointing to reduce performance degradation. Third, we propose a technique to capture consistent snapshots of FPGA and the rest of the computing system. Fourth, we propose a static analysis and a runtime analysis to reduce LUT consumption and performance degradation caused by checkpointing hardware. For host software, we also provide CPRtree stack including API functions to manage checkpoint/restart procedures on FPGA. Our experimental results show that the checkpointing architecture causes up to 9.73% maximum clock frequency degradation, small breakdown, and small data footprint, while the LUT overhead varies from 17.98% (Dijkstra) to 160.67% (Matrix Multiplication). ***

language of the presentation: *** English ***