利用粒子群優化演算法改善分群演算法在訊號分群上之應用 ; To use Particle Swarm Optimization improve K-means Perform Mathematical Calculations in Clustering of Signal Group

NCUIR > College of Electrical Engineering & Computer Science > Executive Master of Communication Engineering > Electronic Thesis & Dissertation > Item 987654321/44534

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/44534

Title:	利用粒子群優化演算法改善分群演算法在訊號分群上之應用;To use Particle Swarm Optimization improve K-means Perform Mathematical Calculations in Clustering of Signal Group
Authors:	黃昆輝;Huang Kun
Contributors:	通訊工程研究所碩士在職專班
Keywords:	分群演算法;利用粒子群優化演算法;模糊決策;群體智能;Swarm Intelligence;Fuzzy Decision;PSO;K-Means
Date:	2010-07-22
Issue Date:	2010-12-09 13:48:01 (UTC+8)
Publisher:	國立中央大學
Abstract:	近年在無線通訊產業的蓬勃發展下，各無線系統也因為技術的突破與量產的實現下，現在無線產品已經廣泛的涵蓋在各種的生活應用上，如現有的手機、電腦、以及新世代研發的數位家電等等，硬體建設上如台北市地方政府建構的WIFLY，跟台北捷運文湖線，以及各學校教育單位在校區內的WIFI，無線系統的應用現在已經大量充斥在生活週遭的各個角落。但是相對的RF頻帶的使用上，在涵蓋的區域及訊號的強弱適當與否，或者甚至於該區域上的監測，皆是各系統在架設起來後需要去維護與管控的。因為各種系統的訊號涵蓋在整個環境的周圍，要如何去評估區分訊號便是本論文深入研究的部分與目的。粒子群優化演算法（Particle Swarm Optimization），簡稱為PSO，此演算法效果具有收斂且運算快速的特性，是解決與應用在進世代類神經網路的研究與最佳化問題上的好方法。PSO在群體內的各粒子間，有著交互訊息溝通的特點，在相關性上可以提供相當適當的資訊，並且具有參數設定少、搜尋速度快、執行實現度高的優點。本論文主題則是以此，強化分群演算法（K-Means）的各粒子在群體間的相關性，提供足夠的權重以適當進行判斷、修正訊號在統計群體內正確的位置，並結合圖資系統呈現視覺化效果來呈現。Because the flourishing development of the wireless communication industry during these years , realization that every wireless system has because of break-through and quantity of technology too , the wireless products have been already extensive use in various life , existing mobile phone, computer, and new to research and develop generation digit of the Electrical home appliances , etc. Build WIFLY constructed of the Taipei local government on the construction of hardware , and Taipei MRT of Wenhu（文湖） Line , and WIFI in many school education , the wireless system application use in a large amount now . But in RF frequency band use , signal power of area contained appropriate or not , or the monitoring on this area , there were every system that needed to maintain and in charge of accusing after erecting . Because various system of signals contain whole around the environment , this thesis further investigates how to go to assess the signal of classify . Particle Swarm Optimization (PSO) , it has convergence and fast operation of characteristic that this performs algorithms , this is a good method on the optimization problem generation that solve and apply to the research of neural networks . Every particle in the colony of PSO , There is mutual information communication characteristic , it can offer appropriate information on relevance , and also parameters are set up less 、 fast to search 、 easy to realized . This thesis theme use this to improve K-Means particles relevance among the colonies , offer enough weights by appropriately judging 、 revision of correct signal position in the colony , and combine map system to appear vision result .
Appears in Collections:	[Executive Master of Communication Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	859	View/Open

社群 sharing

Loading...