非結構式同儕網路(Unstructured peer-to-peer networks)都是使用氾濫搜尋演算法(Flooding search algorithm)找出相關資源，然而過去的研究發現這樣演算法需要大量的儲存量或是產生大量的網路流量。為了解決這個問題，先前研究發現社會群體的概念能改善同儕網路搜尋效能不佳的問題，而進一步發展社群式同儕網路(Social-like p2p network)。然而，學者們也發現在社群式同儕網路有三個缺點：1.缺乏紀錄正確回應結果的節點；2.缺乏語意搜尋機制；3.缺乏保留正確回應節點機制。為了克服這些問題，我們提出一套社群關聯式同儕網路(Social Relationship P2P networks)，以海伯法(Hebbian algorithm)為基礎建立每個節點的社群關聯式權重(Social relation weight)。社群關聯式同儕網路透過社群關聯式權重(Social relation weight)挑選合適回應節點。本研究模擬真實社群式同儕網路的實驗環境。實驗結果證明該社群關聯式同儕網路是有意義改善社群式同儕網路搜尋效率，而且該方法的搜尋效能優於Guntella與NeuroGrid。

Unstructured peer-to-peer (p2p) networks usually employ flooding search algorithms to locate resources. However, past studies have shown that it requires a large storage overhead or generate massive network traffic. To address this issue, previous studies have explored the possibility of building efficient p2p networks by clustering peers into communities based on their social relationships, creating a social-like p2p networks. Previous researchers also indicated that the social-like p2p networks exhibit three problems: 1. Each peer node must record the correct response to peers, 2. lack of support semantic searching, and 3. maintain a knowledge index. To cope with those problems, we propose a social relationship p2p networks that uses a measure based on Hebbian algorithm to create a social relation weight. This social relationship p2p networks uses the social relation weight to select the appropriate response peers. This study also conducts comprehensive trace-driven simulations to evaluate this design. Results show that the proposed social relationship p2p networks significantly improves the social-like peer-to-peer search performance compared with Gnutella and NeuroGrid.