在現實世界的無線感測網路中，我們會收集不同種類的資料，使得傳輸行為變得比收集單一種類資料時更為複雜。在本篇論文中，我們使用tree-based topology，分別為收集不同種類資料的applications建立多棵trees，並將收集的資料傳回各自的root (base station)。所有sensors都採用可調整且可能互不相同的傳輸能量等級。每個sensor接收來自它在資料匯集樹上的children之message，並且將接收到的資料與自身的資料做簡單的data aggregation，再將aggregation後的資料傳給其資料匯集樹上的parent。本篇論文希望在此環境下，針對多個base stations及多個applications的情況來探討如何建立擁有最大生命週期的資料匯集樹集合。
為了達成這個目的，我們的演算法會透過改變資料匯集樹結構的方式，降低normalized load較高的nodes的normalized load，讓所有node的normalized load調整到差不多。降低重度負載節點的normalized load這個動作我們稱為改進 （improvement）。演算法會一直執行到沒有任何符合條件的重度負載節點被改進為止。
我們以傳輸能量等級可調整的model當作實驗組，傳輸能量等級不可調整(固定為最大傳輸能量等級)的model當作對照組，探討在不同node數、不同sensor分布場地範圍大小、不同初始電量範圍、不同Application數這些因素，對lifetime造成的影響。
由simulation的結果我們可以得知，傳輸能量等級可調整的model在大部分情況下，都比不可調整傳輸能量等級的model更能延長無線感測網路的lifetime。

In real-world wireless sensor networks, we collect different kinds of data, making transmission behavior more complicated than collecting a single type of data. In this paper, we use tree-based topology to create multiple trees for applications that collect different kinds of data, and pass the collected data back to their respective root (base station). All sensors use adjustable transmission levels that may differ from each other. Each sensor receives the message from the children in the data aggregation tree, and performs simple data aggregation on the received data and its own data, and then transmits the information after the aggregation to the parent on the data aggregation tree. We hope to explore how to build the data aggregation trees with the maximum lifetime for multiple base stations and multiple applications in this environment.
Our algorithm will reduce the normalized load of the nodes with higher normalized load by changing the data aggregation tree structure, so that the normalized load of all nodes is adjusted to be almost similar. The action of reducing the normalized load of a heavily loaded node is called improvement. The algorithm will continue until there are not any heavy loaded nodes can be improved.
We use the model with adjustable transmission power level as the experimental group, and the model with the transmission power level unadjustable (fixed to the maximum transmission power level) is used as the control group. We use different factors, including region size, initial power range, number of nodes, and number of applications.
From the results of simulation, we can see that the model with adjustable transmission power level can extend the lifetime of the wireless sensor network in most cases than the model with unadjustable transmission power level.

Abstract-----------------------------------------------3
1. Introduction-------------------------------------4
2. Sensor Network Model-----------------------------6
3. Maximum Lifetime Tree Problem--------------------9
4. Algorithm
4.1. Definition of improvement-------------------------12
4.2. Classification of nodes---------------------------14
4.3. Verification of principle-------------------------15
4.4. Detail of algorithm-------------------------------17
4.5. Computational complexity of the algorithm---------19
5. Simulation---------------------------------------22
6. Conclusion---------------------------------------29
References

[1] K. Konstantinos, and S. Tang. “A combinatorial algorithm for the maximum lifetime data gathering with aggregation problem in sensor networks.” Computer Communications, vol. 32, no. 15, pp. 1655-1665, 2009
[2] Y. Xue, Y. Cui, and K. Nahrstedt, “Maximizing lifetime for data aggregation in wireless sensor networks,” Mobile Networks and Applications, vol. 10, no. 6, pp. 853-864, Dec. 2005.
[3] K. Kalpakis, K. Dasgupta, and P. Namjoshi, “Efficient algorithms for maximum lifetime data gathering and aggregation in wireless sensor networks,” Computer Networks, Vol. 42, no. 6, pp. 697-716, Aug. 2003.
[4] H. Ö. Tan and I. Körpeoğlu “Power Efficient Data Gathering and Aggregation in Wireless Sensor Networks,” ACM SIGMOD Record, vol. 32, no. 4, pp. 66-71, Dec. 2003.
[5] Y. Wu, Z. Mao, S. Fahmy, and N. B. Shroff, “Constructing Maximum-Lifetime Data-Gathering Forests in Sensor Networks,” IEEE Transaction on Networking, vol. 18, no. 5, pp. 1571-1584, Oct. 2010.
[6] H. C. Lin, F. J. Li, and K. Y. Wang, “Constructing maximum-lifetime data gathering trees in sensor networks with data aggregation,” in Proceedings of the IEEE ICC, May 2010.
[7] H. C. Lin, W. Y. Chen, "An approximation algorithm for the maximum-lifetime data aggregation tree problem in wireless sensor networks", IEEE Trans. Wireless Communications, vol. 16, no. 6, pp. 3787-3798, Jun. 2017.
[8] A. P. Azad and A. Chockalingam, "Mobile base stations placement and energy aware routing in wireless sensor networks," in Proc. of IEEE WCNC'06, Vol. 1, 3-6 Apr. 2006, pp. 264-269.
[9] S. R. Gandham, M. Dawande, R. Prakash, S. Venkatesan, "Energy efficient schemes for wireless sensor networks with multiple mobile base stations", Proceedings of IEEE Globecom 2003, vol. 1, pp. 377-381, December 1–5 2003.
[10] S. Madden, R. Szewczyk, M. J. Franklin, and D. Culler, “Supporting Aggregate Queries Over Ad-Hoc Wireless Sensor Networks,” In Proceedings of 4th IEEE Workshop on Mobile Computing and Systems Applications, pp. 49-58, June 2002.