﻿ 基于小世界网络的电动汽车市场接受度预测模型
 文章快速检索
 同济大学学报(自然科学版)  2017, Vol. 45 Issue (8): 1160-1166.  DOI: 10.11908/j.issn.0253-374x.2017.08.009 0

### 引用本文

WANG Ning, PAN Huizhong, LIU Xiang, TANG Linhao. Prediction Model of Market Acceptance of Electric Vehicles Based on Small World Network[J]. Journal of Tongji University (Natural Science), 2017, 45(8): 1160-1166. DOI: 10.11908/j.issn.0253-374x.2017.08.009.

### 文章历史

1. 同济大学 汽车学院，上海 200092;
2. 同济大学 交通运输工程学院，上海 200092

Prediction Model of Market Acceptance of Electric Vehicles Based on Small World Network
WANG Ning1, PAN Huizhong1, LIU Xiang2, TANG Linhao1
1. School of Automotive Studies, Tongji University, Shanghai 200092, China;
2. College of Transportation Engineering, Tongji University, Shanghai 200092, China
Abstract: On the basis of the complex social network theory and consumer decision-making theory, a forecasting model of electric vehicle (EV) market acceptance was developed based on the small world network, and the adoption and diffusion process of EVs in social network was simulated through Netlogo software. The simulation results indicate that the EV market acceptance are decided by both the individual preferences and the social network utility and the local network utility show greater influence than the global network utility. Furthermore, selecting the opinion leaders as the initial adopters, increasing the proportion of initial adopters and reducing the purchase threshold of consumers can help to promote the adoption of EVs.
Key words: electric vehicle    social network    consumers' decision-making    market acceptance

1 基于小世界网络的电动汽车市场接受度预测模型 1.1 基于小世界网络的电动汽车市场接受度预测模型构建

1.1.1 模型假设与参数

1.1.2 模型具体分析

(1) 非认知状态—认知状态

${A_{\rm{d}}}\left( t \right) + \sum\limits_{j = 1}^N {{A_{ij}}\left[{{x_j}\left( t \right) + {y_j}\left( t \right) + {z_j}\left( t \right)} \right]} \ge {S_i}, {x_i}\left( {t + 1} \right) = 1$

(2) 认知状态—意愿购买状态

Ui(t)≥Ti时，yi(t+1)=1

 ${U_i}\left( t \right){\rm{ = }}{\alpha _i}{R_i} + {\beta _i}Q\left( t \right) + {\gamma _i}{G_i}\left( t \right), {\alpha _i} + {\beta _i} + {\gamma _i} = 1$ (1)

① 消费者个体初始效用Ri：即是指购买电动汽车能为其带来的效用，是消费者的自身评价，不考虑社会网络的影响.由于个体偏好的差异性，本文参照娄思源[11]的研究，假设消费者个体对电动汽车的初始个体效用Ri服从正态分布，即

 ${R_i} \sim N\left( {{\mu _1}, {\sigma ^2}_1} \right)$ (2)

② 社会网络效用：全局网络效用Q(t)，定义为整个社会网络中购买电动汽车的总人数占比和，如式(3) 所示.局部网络效应Gi(t)即是指消费者个体社交网络中了解电动汽车的其他个体对其电动汽车接受度的影响，本文假设消费者所处的状态越靠后，对社交网络中其他个体的影响越大，假设处于认知状态、意愿购买状态和决定购买状态的影响权重参数分别为μρσ，且μρσ，如式(4) 所示.

③ 消费者的购买意愿阈值Ti：是消费者对购买电动汽车成本的评估.Eppstein等[12]的研究发现收入越高的人在购车时感知风险越小，对电动汽车较高的购置价格不敏感，越有可能成为早期采用者.Slater等[13]通过调查英国的电动汽车车主发现电动汽车的早期购买者不在意较高的购置成本，他们通常家里不止一辆车，收入较高，学历较高，并且有停车位.笔者2015年关于电动汽车市场接受度的研究[14]表明消费者不愿意购买电动汽车的影响因素有充电便利性和感知风险等，即日均出行里程越长的车主越觉得电动汽车出行时充电不方便.因此本文假设购买意愿阈值与消费者的家庭平均年收入成反比，与日均出行里程成正比.

④ 权重参数αiβiγi：表示个体i对不同效应的重视程度，值介于0和1之间.本研究参照张晓军[8]和Delre等[15]的研究，假设所有个体的权重分配都是一样的.

 $Q\left( t \right) = \frac{1}{N}\sum\limits_{j = 1}^N {{z_j}\left( t \right)}$ (3)
 ${G_i}\left( t \right) = \frac{1}{{{k_i}}}\sum\limits_{j = 1}^N {{A_{ij}}\left[{\mu {x_j}\left( t \right) + \rho {y_j}\left( t \right) + \sigma {z_j}\left( t \right)} \right]}$ (4)

(3) 意愿购买状态—决定购买状态

 $R\left( t \right) = \sum\limits_i^N {{z_i}\left( t \right), 0 < t < T}$
2 基于小世界网络的电动汽车市场接受度预测仿真 2.1 仿真系统的构造

 图 1 电动汽车市场接受度仿真流程图 Fig.1 EV market acceptance simulation flow chart
2.2 模型参数初始化

2.2.1 仿真参数

time-ticks：仿真时钟，time-ticks每增加1，程序循环一次，更新消费者个体的状态；

2.2.2 个体属性参数

2.2.3 状态演变参数

(2) 认知阈值Si：假设消费者的认知阈值服从1~3之间的随机分布(消费者若要进入认知阶段，其必须至少有一个影响者).

(3) 个体效用Pi：假设消费者的个体偏好初始值服从正态分布，均值为0.2，方差为0.1，即Pi~N(0.2, 0.1).

(4) 权重参数：权重参数αβγ表示消费者对个体效应、间接网络效应和直接网络效应三者的重视程度，本文假设权重参数αβγ的初始值分别为0.4、0.4和0.2.

(5) 购买意愿阈值Ti：购买意愿阈值与消费者的属性特征有关，基于第1.1.2节分析，发现收入和日均出行距离对消费者的购买意愿影响显著，因此本文假设高收入和日出行距离较短的消费者群体的阈值相对较低，低收入和日出行距离较长的群体阈值较高.具体的阈值水平设定见表 1.

(6) 社会网络的影响程度：个体所处的状态越靠后，其对周边人的影响越大，因此本文假设处于认知状态、意愿购买状态和决定购买状态的个体的影响程度μρσ分别是0.5、0.8和1.

(7) 消费者产生购买意愿之后，以购买概率P=0.5选择购买电动汽车.

2.3 仿真结果

 图 2 小世界网络仿真结果 Fig.2 Simulation result of small world network

(1) 初始采用者比例

 图 3 初始采用者比例与电动汽车市场接受度的关系 Fig.3 Relationship between initial adopters ratio and EV market acceptance

(2) 初始采用者类型

 图 4 初始采用者的类型对电动汽车市场接受度的影响 Fig.4 Influence of initial adopters' type on electric vehicle market acceptance

(3) 权重参数

 图 5 权重参数对电动汽车市场接受度的影响 Fig.5 Influence of weight parameters on EV market acceptance

(4) 个体初始效用的差异

 图 6 初始效用差异对电动汽车市场接受度的影响 Fig.6 Effect of initial utility difference on EV market acceptance

(5) 购买意愿阈值

 图 7 购买意愿阈值对电动汽车市场接受度的影响 Fig.7 Influence of purchase intention threshold on EV market acceptance
3 结论

(1) 电动汽车的市场接受度取决于消费者个体初始偏好和社会网络效应的共同作用，在电动汽车的市场接受度方面，局部网络效应的影响大于全局网络效应.

(2) 电动汽车的最终接受度随初始采用者比例的增加而逐渐增大，并且在初始采用者比例较低(4%以下)时，电动汽车的接受度对初始采用者比例的敏感性较强.

(3) 当随机选取个体作为初始采用者时，电动汽车市场接受度要小于选取意见领袖作为初始采用者的.

(4) 消费者个体对电动汽车的初始偏好差异性越大，电动汽车的市场接受度增长越快，达到稳定状态所需的时间越短，但最终的电动汽车市场接受度却越低.

(5) 随着消费者购买意愿阈值的降低，电动汽车的市场接受度显著提高，且达到稳定状态所需的时间更短.

 [1] 谢玮. 中国原油期货如何吸引国际投资者[J]. 中国经济周刊, 2015(4): 66 XIE Wei. How do China's crude oil futures attract international investors[J]. China Economic Weekly, 2015(4): 66 [2] ROGERS E M, MEDINA U E, RIVERA MA, et al. Complex adaptive systems and the diffusion of innovations[J]. The Innovation Journal: The Public Sector Innovation Journal, 2005, 10(3): 1 [3] AXSEN J and KURANI K S. Interpersonal influence in the early plug-in hybrid market: Observing social interactions with an exploratory multi-method approach[J]. Transportation Research Part D Transport & Environment, 2011, 16(2): 150 [4] TRAN M. Agent-behaviour and network influence on energy innovation diffusion[J]. Communications in Nonlinear Science and Numerical Simulation, 2012, 17(9): 3682 DOI:10.1016/j.cnsns.2012.01.016 [5] KIERMER V. Six degrees of separation[J]. Nature Methods, 2006, 3(1): 1 DOI:10.1038/nmeth0106-1 [6] COX A, GRANBOIS D and SUMMERS J. Planning, search, certainty and satisfaction among durable buyers: a longitudinal study[J]. Advances in Consumer Research, 1983, 10(4): 394 [7] CHO Y, BLOMMESTEIN K V. Investigating the adoption of electric vehicles using agent-based model[C]//Management of the Technology Age. PICMET '15 Conference, Portland:[s.n.], 2015: 2338-2345. [8] 张晓军. 基于复杂网络的创新扩散随机阈值模型研究[D]. 成都: 电子科技大学, 2009. ZHANG Xiaojun. Research on stochastic threshold model in innovation diffusion:a complex network perspective[D]. Chengdu: University of Electronic Science and Technology of China, 2009. http://d.wanfangdata.com.cn/Thesis/Y1505472 [9] MCFADDEN D. Economic choices[J]. American Economic Review, 2001, 91(3): 351 DOI:10.1257/aer.91.3.351 [10] VALENTE T W. Social networks, diffusion processes in[M]. New York: Springer 2009. [11] 娄思源. 基于社会网络的创新扩散仿真研究[D]. 杭州: 浙江大学, 2012. LOU Siyuan. Simulation research on innovation diffusion based on social network[D]. Hangzhou: Zhejiang University, 2012. http://d.wanfangdata.com.cn/Thesis_Y2060176.aspx [12] EPPSTEIN M J, GROVER D K, MARSHALL J S, et al. An agent-based model to study market penetration of plug-in hybrid electric vehicles[J]. Energy Policy, 2011, 39(6): 3789 DOI:10.1016/j.enpol.2011.04.007 [13] SLATER S, DOLMAN M. Strategies for the uptake of electric vehicles and associated infrastructure implications[R]. Cambridge:[s.n.], 2009. [14] 王宁, 晏润林, 刘亚斐. 电动汽车潜在消费者特征识别和市场接受度研究[J]. 中国软科学, 2015(10): 70 WANG Ning, YAN Runlin and LIU Yafei. Identifying consumer characteristics and public acceptance of electric vehicles in China[J]. China Soft Science, 2015(10): 70 DOI:10.3969/j.issn.1002-9753.2015.10.007 [15] DELRE S A, JAGER W, BIJMOLT T H A, et al. Will it spread or not? the effects of social influences and network topology on innovation diffusion[J]. Journal of Product Innovation Management, 2010, 27(2): 267 DOI:10.1111/jpim.2010.27.issue-2 [16] BOHLMANN J D, CALANTONE R J and ZHAO M. The Effects of market network heterogeneity on innovation diffusion: an agent-based modeling approach[J]. Engineering Applications of Artificial Intelligence, 2010, 27(5): 741 [17] Rogers E M. Diffusion of innovations[J]. Isis, 1963, 54(2): 62 [18] GOLDENBERG J, LIBAI B and MULLER E. The chilling effects of network externalities: Perspectives and conclusions[J]. International Journal of Research in Marketing, 2010, 27(1): 22 DOI:10.1016/j.ijresmar.2009.12.005 [19] RAHIM S A. Diffusion model in development: mass media and personal influence[J]. Jurnal Komunikasi Malaysian Journal of Communication, 1985(1): 56 [20] Lehmann K. Innovation diffusion theory[M]. Munich: GRIN, 2004.