電子商務(wù)平臺(tái)中賣(mài)方誠(chéng)信博弈模型研究

紀(jì)淑娟　包春曉　孫冠冠　張純金

摘 要：賣(mài)方為追求利潤(rùn)快速增長(zhǎng)會(huì)通過(guò)各種攻擊策略破壞電子商務(wù)平臺(tái)信譽(yù)系統(tǒng)。現(xiàn)有抵御策略大多是通過(guò)機(jī)器學(xué)習(xí)等相關(guān)技術(shù)從過(guò)程或結(jié)果治理角度實(shí)現(xiàn)，但這些方法很難從根本上解決虛假交易問(wèn)題。針對(duì)以上問(wèn)題，結(jié)合博弈論機(jī)制設(shè)計(jì)領(lǐng)域相關(guān)知識(shí)，從減少造假動(dòng)機(jī)角度出發(fā)，提出一種適應(yīng)于無(wú)壟斷型電商市場(chǎng)對(duì)稱(chēng)演化的賣(mài)方誠(chéng)信博弈模型。模型穩(wěn)定性理論分析和模擬仿真實(shí)驗(yàn)都直觀(guān)清晰地展示出模型中各參數(shù)對(duì)電商市場(chǎng)中賣(mài)家誠(chéng)信演化穩(wěn)定性的影響。實(shí)驗(yàn)結(jié)果表明，從動(dòng)機(jī)角度出發(fā)，采取相應(yīng)措施能夠又好又快地促進(jìn)更多的賣(mài)方選擇誠(chéng)實(shí)交易策略。隨著演化的進(jìn)行，虛假交易將最終退出市場(chǎng)。

關(guān)鍵詞：電子商務(wù);造假動(dòng)機(jī);機(jī)制設(shè)計(jì);演化博弈;無(wú)壟斷

DOI：10. 11907/rjdk. 191467

中圖分類(lèi)號(hào)：TP301 ? 文獻(xiàn)標(biāo)識(shí)碼：A??????????????? 文章編號(hào)：1672-7800（2020）003-0021-06

Research on Sellers Honest Game Model on E-commerce Platform

JI Shu-juan1，2，BAO Chun-xiao1，SUN Guan-guan1，ZHANG Chun-jin3

（1. College of Computer Science and Engineering， Shandong University of Science and Technology;

2. Key Laboratory for Wisdom Mine Information Technology of Shandong Province， Shandong University of Science and Technology;

3. Network Information Center， Shandong University of Science and Technology， Qingdao 266590，China）

Abstract：In order to pursue rapid profit growth， many sellers damage the reputation system of e-commerce platform through various attack strategies. Most of the existing defense strategies are implemented from the perspective of process or result governance through machine learning and other related technologies， but these methods are difficult to fundamentally avoid the occurrence of false transaction phenomenon. In view of the above problems， by combining with the relevant knowledge in the field of game theory mechanism design， this paper puts forward a sellers honesty game model adapted to the symmetrical evolution of non-monopoly e-commerce market from the perspective of reducing the motivation of fraud. Both the theoretical analysis of the model stability and the simulation experiment clearly and intuitively demonstrate the influence of the parameters in the model on the integrity evolution stability of sellers in the e-commerce market. The experimental results show that， from the perspective of motivation， taking appropriate measures can promote more sellers to choose honest trading strategies quickly and well. As the evolution progresses， there will be no more false transactions in the market eventually.

Key Words：E-commerce; fraud motivation; mechanism design; evolutionary game; no monopoly

0 引言

電子商務(wù)平臺(tái)在繁榮景象的背后已經(jīng)出現(xiàn)了非常嚴(yán)重的信任問(wèn)題。相比現(xiàn)實(shí)生活中的面對(duì)面交易，網(wǎng)絡(luò)交易具有匿名性、虛無(wú)性和不對(duì)稱(chēng)性等特點(diǎn)，致使在電子商務(wù)平臺(tái)中建立信任比較困難。

現(xiàn)有的信用評(píng)價(jià)模型能夠防御大多數(shù)攻擊策略，但不能抵御所有信用攻擊。在造假評(píng)價(jià)比重占比較多的情況下，基于累加方法的信用模型（AARE模型[1]、E-Sporas模型[2]和C2CRep模型[3]）對(duì)女巫的抗攻擊性較差;基于過(guò)濾方法的信用模型（iCLUB模型[4-5]和BRS模型[6]）和基于信念的信用模型（LTSL模型[7-8]和TNA-SL模型[9]）能夠抵御大多數(shù)信用攻擊策略，但對(duì)女巫攻擊和漂白攻擊抵御能力較差;基于折扣方法的信用模型（TRAVOS模型[10]和Personalized模型[11]）對(duì)女巫攻擊和共謀攻擊的抗擊能力較差;基于貝葉斯概率的信用模型（FG-Trust模型[12]）雖然對(duì)抵御女巫攻擊和共謀攻擊效果較理想，但對(duì)漂白攻擊的抗擊能力較差;基于進(jìn)化方法的信用模型（MET模型[13]和PEALGA模型[14]）又對(duì)不公平評(píng)分的抗攻擊能力較差。