隨著無線行動通訊系統的發展，為滿足眾多用戶通訊服務品質(QoS)需求，極大化頻譜效率是重要課題，而佈建密集小細胞(dense small cell)被視為有效解決方法之一，因此大、小細胞並存的異質網路(heterogeneous network, HetNet)被廣泛討論。異質網路可提高頻譜效率，但也伴隨嚴重的大小細胞間互干擾問題，所以3GPP Release 10標準制定增強型基地台間干擾協調(enhanced Inter-cell Interference, eICIC)技術以疏解此問題。
eICIC機制包含空白子訊框(Almost Blank Subframe, ABS)與細胞涵蓋展延(Cell Range Expansion, CRE)兩種技術。本論文將針對eICIC機制，以減少大小細胞間的資訊交換的分散式處理為設計主軸，採用人工智慧(Artificial intelligence, AI)中機器學習(Machine Learning, ML)的Fuzzy Q learning (FQL)技術，以通話阻斷率(call block rate, CBR)和通話中斷率(call block rate, CDR)作為關鍵績效指標，研討適應於變動性行動通訊網路的CIO及ABS Ratio設定方法，以提升系統整體服務品質。

To satisfy data-communication demands of mobile communications, through the deployment of small cells can maximize spectrum efficiency. Therefore, the heterogeneous networks (HetNets) have been widely discussed in LTE system. However, because of the serious mutual interference between cells, it will increase the probability of service interruption. Therefore, 3GPP LTE standard Release 10 formulate enhanced inter-cell interference coordination (eICIC) technique to improve this problem.
eICIC include two technology, almost blank subframe (ABS) and cell range expansion (CRE) . This thesis adopt fuzzy q-learning approach and regard distributed process as main idea . We consider call bock rate (CBR) and call drop rate (CDR) as key performance indicators to discuss adaptive adjustment of cell individual offset (CIO) and almost blank subframe ratio (ABS Ratio).

[1] 3GPP TR 36.839, “Mobility enhancements in heterogeneous networks,” 2012.
[2] 3GPP, “Radio Resource Control (RRC)," TS 36.331 V11.2.0, Dec, 2012.
[3]3GPP, “Requirements for support of radio resource management,” TS 36.133 V10.6.0, Mar. 2012.
[4] 3GPP TS 36.300, “Overall description; Stage 2,” 2009.
[5] 3GPP, “Handover procedures,” TS 23.009 V11.2.0, Dec, 2012.
[6] 3GPP TSG RAN WG1 R1-083813, ”Range expansion for efficient support of heterogeneous networks,” Qualcomm Europe, RAN1 #54bis,September 2008.
[7] 3GPP TR 36.814, “Further advancements for E-UTRA physical layer aspects” 2010.
[8] 3GPP TR 36.942, "Radio Frequency (RF) system scenarios," 2012.
[9] 3GPP TR 36.931, “Radio Frequency (RF) requirements for LTE Pico Node B,” 2011.
[10] Fuzzy Q-Learning Process for Enhanced Inter-cell Interference Coordination in LTE Heterogeneous Networks , Liang-You Wang ,Institute of Computer and Communication Engineering National Cheng Kung University, 2019
[11] S. Vasudevan, et al.: Dynamic eICIC-A Proactive Strategy for Improving Spectral Efficiencies of Heterogeneous LTE Cellular Network by Leveraging User Mobility and Traffic Dynamics, IEEE Transactions on Wireless Communications, vol. 12, Issue 10, pp. 4956-4969, Oct. 2013.
[12] J. A. Ayala-Romero, J. J. Alcaraz, J. Vales-Alonso, E. Egea-López, "Online optimization of interference coordination parameters in small cell networks", IEEE Trans. Wireless Commun., vol. 16, no. 10, pp. 6635-6647, Oct. 2017.
[13] S. Deb, P. Monogioudis, J. Miernik, and J. P. Seymour, “Algorithms for enhanced inter-cell interference coordination (eICIC) in LTE HetNets,” IEEE/ACM Trans. Networks, vol. 22, no. 1, pp. 137–150, Feb. 2014.
[14] M. Simseki, M. Bennis, A. Czylwiki, "Dynamic Inter-Cell Interference Coordination in HetNets: A Reinforcement Learning Approach", IEEE Global Communications Conference (GLOBECOM), pp. 5446-5450, December 2012.
[15] A. Daeinabi and K. Sandrasegaran, “A Fuzzy Q-learning Approach for Enhanced Intercell Interference Coordination in LTE-Advanced Heterogeneous Networks,” in Proc. 20th Asia–Pacific Conf. Commun. (APCC), Oct. 2014, pp. 139–144.
[16] Jin Wu; Jing Liu; Zhangpeng Huang; Shuqiang Zheng, “Dynamic Fuzzy Q-Learning for Handover Parameters Optimization in 5G multi-tier networks”, International Conference on Wireless Communications & Signal Processing (WCSP), pp. 1–5, 2015
[17] D. Pandey and P. Pandey, ”Approximate Q-Learning: An Introduction”, Second International Conference on Machine Learning and Computing (ICML), 2010.