2nd International Conference on Advances in Mechanical Engineering and Industrial Inforatics (AMEII 206) Entropy Method based Evaluation for Spectru Usage Efficiency of International Mobile Telecounication Systes* Liqin Wang, Tan Wang 2, Zhaojun Qian2 and Wei Li2* North China Electric Power University, Beijing, 02206, China, 339795680@63.co 2 The State Radio Monitoring Center, Beijing, 00037, China ** Correspoding author: The State Radio Monitoring Center, No.80 Xicheng Disrict, Beijing, 00037, China. liwei@srrc.org.cn. Keywords: International Mobile Telecounication (IMT) Systes, evaluating the usage of frequency bands, evaluation indicators Abstract. Currently the radio spectru resource has becoe increasingly scarce. Evaluating the usage of frequency bands with existing radio services, strengthening anageent of allocated resource and iproving the spectru utilization have been becoing an iportant way of iproving spectru anageent level. In particular, it is urgent to evaluate the efficiency of spectru usage for International Mobile Telecounication (IMT) Systes which is in great spectru deand. In this paper, an evaluation indicator odel is proposed according to the current situation of IMT Systes spectru usage. The set of evaluation indicators is different fro those of the conventional obile counication network which is generally used for the operator network planning and network optiization. The study of evaluation odel in this paper coes fro spectru anager s perspective. Firstly, this paper introduced investigation of spectru anageent at hoe and aboard, then described the proposed evaluating indicators, finally we creatively cobined entropy ethod with saple data based on the odel.it can be seen that the proposed evaluation odel is reasonable and helpful to support relevant spectru anageent work. I. Introduction Since 987, China's obile counication has grown rapidly over the past thirty years which has achieved rearkable developent both in scale and technology. In the end of 205, the total nuber of China's obile phone users reached to.306 billion, net iproving 9.645 illion copared with last year, the obile phone user penetration ratio reached to 95.5%; users service has experienced fro single voice service to the rapidly broadband data service. By 205, obile Internet access traffic consuption reached to 4.87 billion G, iproving 40.% over last year, in which obile Internet traffic reached 3.759 billion G []. The gradual developent of coercial 4G and sart productions has greatly iproved the efficiency of spectru usage [2]. However, doestic and foreign research institutions show that in 2020, with the developent of wireless technology, connecting hundreds of billions of devices and constantly increasing diverse business needs pose a serious challenge for us. the future of obile counications service traffic will show explosive growth. So the evaluation of previous 2G, 3G and even 4G s spectru usage efficiency will be great realistic significance. Evaluation of the frequency bands utilization ais to reflect its status usage efficiency by establishing the indicator syste of IMT. Coprehensive evaluation is widely applied in any fields, such as industry, agriculture, econoy, Coputer and so on which was firstly proposed fro the article "test statistics" published in 888 by Edgeworth [3]. Fro the angle of subjective weighting, people often use expert evaluation ethod, fuzzy coprehensive evaluation ethod, AHP [4] and so on. As for objective approaches, there are principal coponent analysis ethod, entropy ethod, neural network, TOPSIS ethod [5], etc. This paper innovatively uses the entropy ethod to processing data. 206. The authors - Published by Atlantis Press 643
II.Research status at hoe and abroad Countries all attach great iportance to anaging spectru resource because the spectru resource is extreely valuable. They have done a large nuber of work for spectru usage evaluation, spectru anageent work of the typical regions is as follow: The U. S. announced to work for spectru resource inventory between 20 to 205 on April 3th 200 by RADIO Spectru INVENTORY ACT, which provides inventory content: the radio services authorized to operate in each band of frequencies; the identity of each Federal or non-federal user within each such radio service authorized to operate in each band of frequencies; the activities, capabilities, functions, or issions supported by the transitters, end-user terinals or receivers etc [6]. The U.S. FCC released the BROADBAND PROGRESS NOTICE OF INQUIRY [7] report showing the initiative of evaluating today's fixed terrestrial broadband service-, and put obile and satellite broadband service into the scope of broadband evaluation for the first tie. This report claied that the evaluating indicator ust include speed, latency, consistent, and the data would be obtained fro several authoritative laboratories such as Ookla, Rootetrics and Google M-Lab for further ipleent of the evaluation. The EU established a project group consisting of 27 eber countries which is responsible for collecting the public and the dedicated spectru applications at 400MHz-6GHz band. The tea proposed a technology odel and social benefit evaluation syste of spectru application to assess the technical efficiency of existing spectru, they copared with other countries in the world for iproving the usage efficiency of specific frequency bands. British the earliest investigated obile counication base station eissions and power level values in 2004 in the report Mobile Phone Base-Station Audit ; British has a long-standing coitent to the audit of spectru fro 2002 with the ai of releasing the axiu aount of spectru to the arket and increasing opportunities for the developent of innovative new services and copleted Independent Audit of Spectru Holdings in 2005[8]. Key recoendations include: the introduction of arket echaniss into spectru anageent in the public sector; changes to the structure and scope of Adinistered Incentive Pricing as applied to the public sector; etc. In suary, countries all coit to spectru anageent and spectru onitoring work in long-ter and the spectru of the inventory is ephasized differently. Soe countries focus on spectru optiization, others focus on spectru recycle. At present, there is not a country using the coprehensive evaluation ethod to evaluate the spectru usage. III.Evaluation index Traditional easure of spectru usage ainly through the spectru occupancy, spectru occupancy is the ost direct way to reflect spectru usage, within a period of tie, if the received signal power strength is higher than the threshold value, which eans the frequency band is occupied, or that is not occupied. In this paper, the evaluation indicator syste is ainly considered fro the three aspects: frequency, service, base stations, which akes up the deficiency of the existing single research ethod. The frequency includes the bandwidth and frequency occupied by the operator, and the service includes the rate of dropped calls, regional average telephone traffic and regional average data rate. The station includes the nuber of base stations within an unit area. The selection of the evaluation indicator is ore coprehensive than the previous, and the correlation between the indicators is sall. 644
Table Public obile counication spectru utilization efficiency evaluation index Index Forula Unit Frequency Bandwidth Allocated Bandwidth [MHz] User nuber The nuber of total user / [nuber/mhz] bandwidth Frequency occupancy devices easure [percentage] Service Drop rate Nuber of dropped calls / [percentage] Total nuber of calls Regional average voice traffic Regional all day traffic/ [Erl/MHz*h] (Allocated Bandwidth*24) Regional average data rate Area throughput/ Allocated [Mbps/MHz] Bandwidth Base Station Nuber of stations Nuber of base stations / (9 square kiloeters) [nuber/ square kiloetre] IV.Evaluation ethod based on Entropy Mobile spectru usage efficiency evaluation syste has a nuber of indicators, how to deterine the weights of these indicators is critical, the entropy ethod is a copletely objective ethod for obtained the indicators weight, this paper will use entropy ethod to evaluate. When there is a great difference between the data of one indicator, that less entropy, indicating that the large effective inforation content as well as large indicator weight; on the contrary, if a target saple data has a saller difference, entropy will be bigger, which indicates the saller aount of inforation and saller weight. So according to the degree of difference between indicators achieving dynaic adjust the weight given by entropy. With objects, n evaluation indicators, the original indicator data atrix is obtained as follows: x x2... x x2 x22... x 2 X = = (X,X 2,...,X ) ()............ xn xn2... x4 If Xi{ i =,2,3... } is a negative indicator, that is, the saller the value, the better indicators, We need to convert negative indicator to positive indicator as (2): X i = Xi{ i =,2,3... } (2) The proportion of the i th saple under the j th indicator is as: x ij pij =,( i =,2,..., n; j =,2,... ) (3) n x ij The entropy the j th indicator as (4): n e = k p In( p ) (4) j ij ij i= Where: k > 0, k = / In( n), e j 0 The difference coefficient of the j th indicator is like the forula (5). e j g j =, e ej E E = (5) e the weight of the j th indicator is as (6). 645
gi ωi = ( j ) g i The saple score is (7): si = ωi pij,( i =,2,... n) (7) (6) V.Instance verification Table 3 is a set of saple data obtained fro the actual anufacturers, the data with entropy ethod obtained the weight of each indicator: w = (0.587, 0.796, 0.20, 0.394, 0.360, 0.204, 0.457), and copanies A, B, C evaluation scores respectively are 0.4023, 0.3285, 0.2692, thus draws the highest spectru efficiency is anufacturer a, followed by anufacturer B, spectru efficiency lowest is anufacturer C. Table 2 saple data saple data anufacturer A anufacturer B anufacturer C bandwidth 40 60 70 User nuber 3.4e8 0.64e8 0.49e8 Frequency occupancy 96% 98% 99% Drop rate 2% 2% 3% Average voice traffic areas 4.46 5.56 4.76 Regional average data rate 00 80 60 Base Station 0e4 46e4 30e4 VI.Conclusion This paper considers that evaluating the usage efficiency of obile counication evaluation should be a process fro junior to senior in a long period of tie. With the continuous iproveent of testing equipent we can obtain ore indicator data for ore coprehensive evaluation; at the sae tie, we can also set indicator syste fro the econoic, social benefits, environental protection aspects etc by future ore advanced eans of signal identification and the analysis of big data for ore scientific and effective evaluation. In this paper, we used entropy ethod to analysis data, in follow-up study we will consider using other ethods showing the result of evaluation. Acknowledgeents * This paper is supported by Chinese National Key Project under Grant No. 205ZX03002008, and National HighTechnology Research and Developent Progra ( 863Progra) of China under Grant No. 205AA0A705 and No.204AA0A706. References [] Inforation on http://www.iit.gov.cn [2] Chen, Qian Zhang, Mingyan Liu, Shufang Li. Mining, in: Spectru Usage Data: A Large-Scale Spectru Measureent Study, IEEE Transactions on Mobile Coputing (202) [3] T.J. Stewart, in: A critical survey on the status of ultiple criteria decision, Oega (99) 646
[4] B. Michele, C. Christine, in: The analytic hierarchy process and the theory of easureent, Manageent Science (200) [5]Xu Xiaozhan, in: A note on the subjective and objective integrated approach to deterine attribute weights, European Journal of Operational Research (2004) [6] Inforation on https://www.ntia.doc.gov [7] Inforation on https://www.fcc.gov/ [8]Inforation on https:http://www.spectruaudit.org.uk 647