MAST: an Agent Framework to Support B2C E-Commerce Salvatore Garruzzo, Domenico Rosaci and Giuseppe M.L. Sarné DIMET - University Mediterranea of Reggio Calabria Località Feo di Vito - 89060 Reggio Calabria, Italy Email: {salvatore.garruzzo, domenico.rosaci, sarne}@unirc.it Abstract— In this paper we present an XML-based multi-agent activities during all the CBB stages. The MAST framework system, called Multi Agent System for Traders (MAST), that presents the following important features: (i) software agents completely supports Business-to-Customer E-commerce activi- are XML-based to manage agent profiles and messages in a ties, including advertisements and payments. MAST helps both customers and merchants in their tasks with a homogeneous light and easy manner and to realize agent communications in and personalized approach. In particular, E-payments in MAST ACML language [3], [7] assuring portability; (ii) an Ontology are implemented under the availability of financial institutions. is used as a common language for all agents allows to unify the This avoids exchanging of sensible customers’ information and representation of products and categories belonging to various reinforces the confidence between customers and merchants. catalogues; (iii) an e-payment protocol called AIPP (Agent A complete prototype of MAST has been implemented in the JADE framework, and it has been exploited for realizing some Internet Payment Protocol) [6], based on existing financial experiments, in order to evaluate its performances. institutions, is fully compliant with the standard FAST [2] and it is used together with single-use account identifiers [18]; (iv) a central agency provides agents with some services and I. I NTRODUCTION cooperates with them to realize only the “Need Identification” A great contribution to the Internet diffusion has been pro- and the “Service and Evaluation” stages of the CBB model in vided by E-Commerce (EC) activities, i.e. all trading activities an efficient way. carried out by means of Internet. In [15] a classification of the The paper is organized as follows: the MAST framework EC activities in homogeneous categories has been realized on is presented in the following section. In Sect. III the AIPP the basis of the typology of the traders and of the specific trade protocol is briefly illustrated and in Sect. IV the adopted func- activity carried out over the Internet. In this paper we deal tionalities for customer and merchant support are described. In with one of these categories, i.e. the Business-to-Consumer Sect. V the MAST prototype and performances are discussed. (B2C), that can be compared to the retail trade of traditional Section VI deals with some Related Work and finally, in commerce. Sect. VII, some conclusions are drawn. Nowadays, the B2C involves a large number of merchants interested in offering products using a convenient media and II. T HE MAST F RAMEWORK customers that desire to purchase those products. In this con- text, customers and merchants can exploit different opportuni- In the MAST framework, represented in Fig. 1, each cus- ties as: (i) absence of time and space boundaries; (ii) simple, tomer C and merchant M is associated to her/his personal fast and comfortable purchases; (iii) low costs and several agent (resp., c and m) and with her/his financial institution sale terms available. However, a significant customer-merchant (F I). All agents are logged into the MAST Agency (Ag). Both distrust still persists, mostly due to the absence of personal agents and agency support B2C activities managing (in terms contacts and to a low acceptance of the e-payment methods of insertion, deletion and updating) their respective Knowledge for security reasons. To capture the different phases carried profiles. In this section, agents and agency will be briefly out by enacting a B2C process, some behavioural models described by illustrating their profiles and behaviours, while can be exploited, and particularly, in this paper we adopt the B2C support activities in the CBB stage are exposed in the Consumer Buying Behavior (CBB) model [8], where the Sect. IV. trading activities have been embedded in six different phases (resp., “Need Identification”, “Product Brokering”, “Merchant A. The MAST Agents Brokering”, “Negotiation”, “Purchase and Delivery”, “Service and Evaluation”). In the following, U denotes the generic user (a customer or a This work presents a multi-agent framework to support B2C merchant) and a represents her/his agent. Each MAST agent activities of merchants and customers. Such a framework, manages its Agent Knowledge (AK) profile, represented in called Multi-Agent System for Traders (MAST), is composed Fig. 2 and described by the following elements: of a set of agents and a central agency. In particular, in MAST • U D (User Data), contains the user’s name (N ame) each merchant and each customer is provided by a software and address (Address), login identifier (AcL), password agent, managing a personal profile, able to support B2C (AcP ), real (Ac) and single-use (AcT ) user’s account 76 following elements: visit counter (V C); first visit (F V ), one before the last visit (P V ) and last visit (LV ) dates; product rate (R); a set (P ASet), where each element is associated to an agent that has been interested in the product, and that is composed by an agent identifier (aI), the highest CBB stage reached and eventually the delivery identifier (DI) and the auction flag (A). More in detail, the rate R represents the interest of a customer for a specific product and it is updated (by a) when a CBB activity is monitored by a using the following formula: P5 CV ·(1+P V −F V ) R = φ · ξ=1 [ ((6−ξ) 2 ·(1+LV −P V )) ] where: ξ ∈ [1, · · · , 5] identifies one among the first five Fig. 1. The MAST architecture CBB stages; φ ∈ {1; −1} describes the satisfaction of the customer about a product (φ is usually set to 1, but it is set to −1 by the customer only if in the last and optional CBB stage he/she is unsatisfied of the purchased product; identifiers, all referred to the user’s account into F I. Note for a merchant, φ is always set to 1). Furthermore, the that Ac and AcT include also F I coordinates. differences among F V , P V and LV are expressed in • AD (Agent Data), containing the user’s agent (aI) and days1 beginning from F V . Note that R depends on the Agency (AgI) identifiers and a pruning threshold (T ) to number of times that an activity has occurred in a CBB delete uninteresting information from AK. stage, the relevance of the involved CBB stage and the • O (Ontology). In order to identify products of interest more recent accesses. for the users, in our framework as Ontology we adopt The information in the described structures are used by an the North America Industry Classification (NAICS) [13], agent a to realize its goals, as explained in the following, which is an official hierarchical industrial classification excluding the CBB support which is presented in Sect. IV. used in North America, employing unified evaluation More in detail: criteria. In MAST, the 6 digits NAICS code is used to • setup steps: semi-automatic procedures are activated to: identify a product. It is clear that other ontologies of this kind might as well be adopted instead of the NAICS (i) set initially or update U D, AD, ADB, interacting coding. with Ag when it is needed as in the first a’s activation; • P D (Product Data) is a set of products, where each (ii) remove a from the system for an U ’s request to Ag. • operational steps: a customer agent is automatically product is represented by an identifier (P I) and described by the following elements: NAICS code (N ); model (M ); activated (resp., deactivated) when a Web session starts brand (B); price (P ); currency (C); commercial unit (U ); (resp., ends “per se” or for an explicit customer’s choice), auction flag (A) that is set to 1 if the product has a fixed whereas a merchant agent is automatically activated price otherwise it is set to 0; tax (X); benefits set (BSet), (resp., deactivated) when its site is on-line (resp., off-line eventually empty, of added values; delivery set (DSet), or for an explicit merchant’s choice). An agent performs that collects the delivery identifiers (see DD section) the following activities: (i) it sends periodically to Ag related to the product. its aA; (ii) it constructs its profile to support its user • DD (Delivery Data) is a set of elements, where each updating its AK w.r.t. each agent contact and each access element, represented by an identifier (DI), is described for a product in one or more CBB stages; (iii) in order by the delivery time (DT ) and by the fixed (F ) and to realize the first phase of the “Need Identification” variable (V ) costs. Note that DD collects the data of the CBB stage, in MAST a customer agent periodically sends chosen delivery for a customer, while it collects the data to Ag a list (L) containing the NAICS code of those of the delivery he/she makes available for a merchant. products that meet interests and preferences of its user, • ADB (Agent Data Base) is a set of agent (and Agency) ordered on the basis of their rate R; (iv) periodically each data where each element, represented by its identifier aI, agent prunes its AK from some evaluated unimportant is described by its Internet address (aA) and by the date information on the basis of the values of the rating w.r.t. of its update (aAU ). the threshold T . • U P (User Profile) is a set of data that an agent a obtains monitoring the CBB activities in the MAST environment; B. The MAST Agency for a customer its c agent collects the data of the products The Agency Knowledge (AgK) profile is described by: which the customer is interested in; elsewhere, for a • AgD (Agency Data), that is composed by the Agency merchant its m agent collects the data of the CBB ac- Identifier (AgI) and Internet Address (AgA). tivities carried out in the site by the agents of the various 1 The choice of the day as reference time unit is due to the characteristics customers for the products offered by the merchant. Each of the problem, given that purchases usually do not occur often in time (e.g., element of U P is represented by the identifier P I (the each minute or hour). However, it is possible to change the reference time same of the P D section) and it is described by the unit without influencing the generality of the model. 77 informs the community. • service managing steps: The Agency provides some services to agents, namely: (i) the support to realize some CBB stages efficiently as will be described in the following; (ii) a broadcasting message service (e.g., to provide an agent m offer to all agents c); (iii) a yellow page service, where each affiliate can ask the address of another MAST affiliate. III. T HE AGENT I NTERNET PAYMENT P ROTOCOL (AIPP) Payment schemes can be assessable with subjective criteria, as customer acceptance or trust [14], and objective criteria, as functionality and quality parameters like transaction cost, security, privacy, etc. The presence of a network in a payment scheme introduces new issues, absent in traditional scenarios [1], [16], where: (i) identities of the transaction actors need to be authenticated and validated; (ii) payments and their effects guaranteed; (iii) operations, frauds and legal risks minimized. In addition, an extended use of standard protocols, existing products and services, payer anonymity, purchases confidentiality and low costs are desirable. Currently, the most used e-payment system is the credit card, but in this case a credit card number should be provided to the merchant; this could be risky because the card num- ber is provided over Internet and/or stored in the merchant site. The electronic cash systems cannot be used due to law and crime prevention regulation/legislation [1]. Recently, centralized account schemes have grew quickly in popularity for their aptitude to integrate usual financial instruments in a secure Internet transaction context. This payment family, also proposed by well known financial institutions, includes general purpose or e-commerce specific applications and can be realized completely either in secure software or in secure hardware. A centralized account approach has been proposed in 1999 by the Financial Service Technology Consortium (FSTC) with Fig. 2. The Agent Knowledge (AK) the Financial Agents Secure Transaction (FAST) project [5]. The FAST team has developed five payment schemes for different scenarios (without specifying any detailed protocol) • ADB (Agent Data Base), that is a set of agent data based on financial institutions that manage user’s accounts where each element, represented by its identifier aI, is and agent technologies to take advantage from existing infras- described by: an Internet address (aA); the date of the aA tructures. The main benefits are: (i) customers and merchants update (aAU ); a list (L) of NAICS code referred to those with no common authentication mechanisms (FAST is not an products of interest or preferences; the name (N ame) and authentication model) are reciprocally authenticated by their address (Address) of the agent’s owner. financial institutions when they log in their on-line accounts • aP T (Agent Pruning Threshold) that is exploited to with the usual procedures (commonly with login and password deallocate long-time inactive agents. over an SSL connection [4]); (ii) payments occur directly via The behaviour of the Agency consists of: financial institutions to guarantee effective funds availability, • affiliate managing steps: The Agency carries out the fol- funds transfer and connected effects, but also promote credit- lowing operations automatically: (i) when it is required, push; (iii) interoperability among accounts located in different the Agency affiliates an agent sending it its Identifier and financial institutions is easy to effect (as between two banks) at this point the agent is logged and operative; (ii) the choosing among different transfer modalities usually available. Agency updates the agent data when it changes; (iii) the On the contrary, it is hard when accounts are located into Agency stores, for each active agent, the current address competitor payment systems; (iv) payments are carried out and the list L that the agent periodically sends to Ag; by agents that replace customers and merchants in most (iv) if a user requires an agent deletion to the Agency or uninteresting and/or complex tasks. if an agent is inactive for a time longer than the pruning The risks in FAST can be further minimized by transferring threshold aP T , then the Agency deletes the agent and funds over interbanking networks, assigning to each message a 78 TABLE I time to live and a unique identifier, managing as much sensible M ESSAGE S PECIFICATION information as possible off-line, etc. The problems of security communication among financial institutions, as those related Message Message Content IN Fx,y H(aS x , aRx , ncx , T T Lx ), to defense against viruses or hacker attacks, are beyond the G(N, M, B, P, C, U, A, X, BSet, DD, CU R, F P ) H(aS c , aRc , ncc , T T Lc ), FAST project objectives. REQ IN Vc,m G(P I M , N, M, B, P, C, U, A, X, BSet, DI M , CU R, F P ) In this paper, we exploit the e-payment protocol AIPP IN Vm,c H(aS m , aRm , ncm , T T Lm , P II m ), G(P I M , N, M, B, P, C, U, A, X, BSet, DI M , DD, F, V, CU R, F P ), F(F II M , (Agent Internet Payment Protocol) [6], complying with the F IAM , AcT M ) P Oc,m H(aS c , aRc , ncc , T T Lc , P II m ), FAST “pre-negotiation” scheme [2], together with single-use F(F II C , F IAC , AcT C , AddressC ) account identifiers [18]. In AIPP a low amount of information P Ex,y H(aS x , aRx , ncx , T T Lx , P II m ) P Ax,y H(aS x , aRx , ncx , T T Lx , P II m ) is exchanged without any explicit encryption level. Moreover, MTO c,F I C H(aS c , aRc , ncc , T T Lc , P II m ), F(F II M , F IAM , AcT M , H(IN Vm,c )) AIPP adopts only asynchronous agent communications with- C C C C A MTO H(aS F I , aRF I , ncF I , T T LF I , P II m ) out multiple Internet connections (other parties connected to F I C ,c C C C C R MTO H(aS F I , aRF I , ncF I , T T LF I , P II m ) the infrastructure, such as Internet providers, are considered as F I C ,c ACT CODx,y H(aS x , aRx , ncx , T T Lx , P II m ), F(H(IN Vm,c )) external risk factors). In this way, it is proposed as a potentially N EW AIx,y H(aS x , aRx , ncx , T T Lx ), F(AcT y ) EV ALc,m H(aS c , aRc , ncc , T T Lc , P II m ) well acceptable Internet financial transaction method able to . In the first three CBB stage the messages can be addressed to c agents chosen among satisfy all issues of an e-payment scheme that have been those listed in ADB or employing the Ag’s broadcasting messages service previously described. • P Ex,y (data) (resp., P Ax,y (data)): it notifies that the IV. T HE MAST S UPPORT TO CBB ACTIVITIES payment has been performed (resp., aborted) w.r.t. P Oc,m (data); MAST provides a support, in accordance to the CBB • M T Oc,F I C (data): it is an irrevocable money transfer model, to customers and merchants in their EC activities. In order w.r.t. IN Vm,c (data); MAST, typical interaction between agents involves a customer • A M T Oc,F I C (data) (resp., R M T Oc,F I C (data)): it (C) with her/his agent (c) and financial institution (F I C ), a notifies the MTO acceptance (resp., rejection) w.r.t. merchant (M ) with her/his agent (m) and financial institution M T Oc,F I C (data); (F I M ), the Agency (Ag), a product (G) offered by a M . • ACT CODx,y (data): it contains the MTO activation The appropriate financial institution typologies are limited to code w.r.t. IN Vm,c (data); banks, card issuers or relevant financial organizations; further, • N EW AIx,y (data): it contains a new single-use account it is assumed that payers and payees can manage their on-line identifier to be employed in the next purchase or sell; accounts. In the following, the terms product and service will • EV ALc,m (data): it is an optional evaluation of a pur- be used interchangeable. chase. In MAST, to avoid possible attacks, single-use account iden- A data XML document is structured in three sections tifiers (preserving also financial privacy) and a nonce (i.e., an including: agent sender marker) are adopted, and a Time To Live (T T L) 1) H (Header) that is composed by: agent identifiers of is used as message deadline for each agent communication. Sender (aS) and Receiver (aR); CBB Stage (S); Nonce Moreover, to promote trust among customers and merchants, (nc) that is an agent’s marker; Time To Live (T T L); the AIPP protocol allows the F Is to be third parties in a Product Invoice Identifier (P II). financial transaction, still guaranteeing user’s privacy. 2) G (Products) that encodes: Product Identifier (P I) and Notation and data contents of the messages used in MAST the product data (N, M, B, P, C, U, A, X, BSet) previ- to transfer in a consistent and efficient way the business infor- ously described in Sect. II-A; one or more Delivery Iden- mation are illustrated before describing the MAST protocol. tifiers (DI) with the corresponding data (DD, F, V ), Note that the subscripts identify sender and receiver while data previously described in Sect. II-A; Commercial Unit is an XML document2, whose content is context sensitive (see Required (CU R); Final Price (F P ). Table I). More in detail: 3) F (Financial) is constituted by: Financial Institution • IN Fx,y (data): it requires/provides commercial informa- Identifier (F II); Financial Institution Internet Address tion about a product; (F IA); Financial Institution Single-Use Account identi- • REQ IN Vc,m (data): it requires an invoice for a product fier (Ac); User Address (Address). offered by M ; The actions performed by agents in MAST to support • IN Vm,c (data): it contains the invoice required with customers and merchants in their B2C activities during all REQ IN Vc,m (data); CBB stages are described below in detail for each CBB stage • P Oc,m (data): it is the purchase order w.r.t. and represented in Fig. 3. IN Vm,c (data); a) Need Identification Support: (ξ = 1). In the first CBB stage, customers identify their needs and merchants advertise 2 MAST agents employ the eXtensible Markup Language (XML) [22] to their offered products (G) to as more potential customers as overcome several heterogeneity problems (platforms, languages, applications possible. In detail: (i) when an M wants to make an offer and communication modalities) and to transfer business information in a consistent way. Note that specific agreements must be established on the tag about a product to potential Cs, he/she has to submit own semantic. offer sends to Ag an IN Fm,c ; (ii) in a first phase the Ag, on 79 the basis of the lists L provided by the c agents, takes care of sending the offer to the potentially appropriate c agents, then in a second phase the c agents present such offers to their Cs only if they are fully compatible with their interests and preferences. b) Product Brokering Support: (ξ = 2). This stage occurs when a customer has identified a need and looks for a suitable product to satisfy it. In detail: (i) C can ask information on the desired product typology to one or more M s by means of IN Fc,m ; (ii) all M s that have a product that matches the C request, reply with a new IN Fm,c with all the details of the products and commercial information. c) Merchant Brokering Support: (ξ = 3). A customer identifies the most suitable merchant to purchase a product carrying out the following actions: (i) If C has sufficient knowledge of the product details, a c’s message IN Fc,m is sent to one or more merchants; (ii) if there is a product that matches C’s request, M replies with a message that reports a complete description of the product; in such a way c can select the best product offer. Note that if in the previous stage C has received a sufficient number of IN Fm,c it is possible to choose a merchant without carrying out this present stage explicitly. d) Negotiation Support: (ξ = 4). In this stage a pair of customer and merchant define the purchase details. They realize suitable strategies in a multi-round session for their respective bids and offers presented by means of messages. This stage is closed when an agreement is reached or the timeout T T L of the last message has elapsed. e) Purchase and Delivery Support: (ξ = 5). In this stage the customer purchases, pays and chooses a delivery modality for a product offered by a merchant employing the AIPP protocol where: payer and payee identities are authen- ticated by their respective financial institutions during their on-line accounts accesses (usually with login and password over a SSL Internet session); payments occur directly among the financial institutions; Single-use account identifiers are adopted; No heavy protocol is needed; no sensible financial and commercial information is exchanged to assure privacy; financial institutions are third parties in the transaction to guarantee customers and merchants. The actions performed Fig. 3. UML of the MAST support activities in this stage are: (i) When C wants to purchase a product offered by M , he/she sends the message REQ IN Vc,m ; (ii) m replies with IN Vm,c (a pro-forma invoice); (iii) c logs into then F I M informs m with a P EF I M ,m message, otherwise F I C and then orders a Money Transfer Order (M T Oc,F I C ) to after the TTL of the ACT CODF I M ,F I C message F I M F I M payee; (iv) F I C accepts/rejects the MTO on the basis informs m with a P AF I M ,m of the sell failure; (x) finally, of the existence of sufficient C’s funds and notifies to c its m could however accept the payment informing F I M , and choice with a A M T OF I C ,c + a new single-use account consequently F I C , or refuse it aborting the sale and returns identifier (AcT ) for the next purchase or with a back the money to F I C by means of its F I M . At last F I C R M T OF I C ,c message; (v) c sends a P Oc,m to effect the will inform c whether the product has been purchased or not. purchase order; (vi) m logs into F I M and sends the required f) Service and Evaluation: (ξ = 6). It is an optional payment activation code (H(IN Vm,c ) to F I M ; (vii) F I M pro- feedback provided by a customer to express her/his dissat- vides M with a new single-use account identifier (AcT ) for the isfaction about the purchase of a product, the merchant or next sell and sends to F I C the payment code (H(IN Vm,c ); both. Two kinds of actions can be carried out by the customer: (viii) if the activation code is the same as that provided by c, (i) if the purchased product has been evaluated negatively, the then F I C effects the payment via F I M and informs c about Rate R ∈ AK will assume a negative value by setting the the state of success (P EF I C ,c ) or failure (P AF II C ,c ) of the φ coefficient to −1; (ii) if the merchant has been evaluated MTO process; (ix) if the payment has been performed by F I C , negatively, its identifier will be deleted from P ASet (w.r.t. G), 80 Time employed by the merchant server for one MAST activity runs at a time. Conversely, a merchant 1000 "Purchase and Delivery" transactions agent is associated with high computational cost, given that it has to satisfy a large volume of processes at the same time referring to different c agents. Consider that the server has to 600 carry out other tasks for its EC activities that can absorb also a 500 significant amount of resources and that have been simulated 400 by assuming some different application costs as percentage of Seconds all processes carried out by an m agent; besides, the Internet 300 cost can influence the global system performance and it has 200 been simulated by setting some delays in the communications 100 (tests were carried out on a 100 Mb LAN). 60 Application This procedure is surely a critical test activity, for the 0 0 50 100 150 0 Cost (%) necessity to coordinate more parties, sending more messages 200 250 and realizing some secure connections (SSL connections have Network Cost (msec) been adopted in the tests). In the following, we employ this process only in order to obtain a rough estimation of the Fig. 4. The payment performances MAST computational efficiency in this CBB stage (keeping in mind that the MAST approach has anyway other significant benefits provided by the authentication mechanism and pay- then C might decide to inform M directly, or in an anonymous ment security level). In particular, the time necessary by the form by means of Ag, about her/his dissatisfaction. merchant’s server to complete a sequence of 1000 “Purchase and Delivery” transactions for different application costs and V. S YSTEM P ROTOTYPE AND E XPERIMENTS network delays is represented in Fig. 4. Note that some steps occur using no secure connections and other occur on a A complete prototype of MAST framework has been im- simulated reserved banking channel. plemented in JADE [10], to test its CBB support activities simulating either single or continued sequences of CBB pro- The experiments suggest some considerations about the cesses in a small B2C scenario3 . Furthermore, to realize such implemented MAST prototype and experimental results ob- experiments some EC sites have been realized in XML. In par- tained. Using the MAST prototype, all CBB activities have ticular, in this section the results of the “Need Identification” been correctly carried out, customer and merchant profiles and “Purchase and Delivery” stages are reported. have been initialized, correctly updated and all our project In the “Need Identification” the experiments have been goals have been meet, showing the capability of MAST to finalized to measure the customer satisfaction degree (CSD) provide proper support to customers and merchants in EC computed as the number of merchants’ offers evaluated by scenarios. The experimental results obtained, even though they the customers as correctly filtered by the system. The filtering have only an indicative meaning both for the initial scenario process is carried out for each customer in two phases, the assumption and some compulsory rough simulation show first performed by the Agency and finalized to disregard im- interesting performances in terms of efficiency, effectiveness mediately all merchants’ offers clearly out from the customer’s and time employed. interests and preferences; while the second phase is performed by the customer’s agent to realize a fine tuning of the filtering VI. R ELATED W ORK activity on the basis of its profile. The tests have been carried The various aspects connected to the B2C have been dealt out by 19 customers, using their agent profiles previously built with in a very large variety of scientific works; some works on the basis of CBB activities carried out on XML EC sites. which to our knowledge come closest to the material presented More in detail, in the first phase on a total of 4750 merchant in this paper will be mentioned in this section. offers (250 offers for each customer), randomly chosen among The role of software agents in the EC has become very the products offered in the XML EC sites, the agency has relevant, as proved by the large number of models and correctly rejected 3154 of them and considered potentially architectures proposed in literature and the state of the art interesting 1596 offers. Then in the second phase, on the has been investigated in a significant number of surveys remaining 1596 offers the customers’ agents have evaluated [8], [9], [11], [12], [15], [19], [23]. The main opportunity surely interesting only 193 offers, but 79 of them were not offered by multi-agent systems is to support customers and really interesting for the customers. In this way, we obtain a merchants in performing their B2C activities. In the CBB global CSD equal to 0,983. Note that we have set the filtering context, MASs were traditionally focused on only in a few parameters to avoid the rejecting useful merchants offers. stages, usually “Merchant Brokering” and/or “Negotiation” About the “Purchase and Delivery” stage, it is clear that stages, but progressively their support has been extended to all the cost located on the customers’ client side has a minimal CBB stages (note that many MASs for B2C do not explicitly impact on the computational performance, since usually only use the CBB model, but their activities are easily brought back 3 The simulations have been realized by employing computers based on to it). Furthermore, only a restricted number of MASs adopt single CPU (Intel Pentium 4, 3 GHz), RAM 2 Gbyte and O.S. Linux. one or more existent payment schemes explicitly, while the 81 largest part of them just ignore the issue or record that a As for ongoing research, a development of MAST is payment has occurred. Finally, since there is a large variety of planned by the introduction of different behavioural models protocols and communication languages that MASs adopt in taking in account emerging behaviours in the B2C area, such B2C, these will not be specifically addressed here. In particular as formation of coalitions or the EC-site visiting. we propose the following three approaches: • MAGMA [20] proposes a MAS for a free-market ar- R EFERENCES chitecture based on messages. In MAGMA the agents [1] Asokan N., Janson P., Steiner M. and Waidner M. The State of the Art in are monitored by a central administration, only par- Electronic Payment Systems. IEEE Computer Magazine, 30(9):28–35, tially automatized; agents provide some trading strategies, September 1997. [2] Financial Services Technology Consortium. Financial Agent Se- independently by the users’ behaviour, and can form cure Transaction (FAST), Phase 1 Final Report (White Paper). agent alliances. Financial services for EC activities are FAST Phase 1 White Paper.pdf, September 2000. provided, in a secure way, by a virtual bank that manages [3] Foundation for Intelligent Physical Agents (FIPA). FIPA ACL Message Structure Specification. SC00061G.pdf, December 2005. specific agents’ account. [4] Freier A.O., Karlton P. and Kocher P.C. The SSL Protocol version 3. • CASBA [21], resulting from a CEE ESPRIT Project, im- ssl-toc.html, November 1996. plements an Internet agent-based marketplace supporting [5] The Financial Services Technology Consortium (FSTC) website, 2006. [6] Garruzzo S., Sarné G.M.L. and Palopoli L. AIPP: A FAST-Complied all CBB stages, in a flexible way with different auction E-Payment Protocol. In IADIS International Conference - Applied types and dynamic negotiations, and some commercial Computing 2006 (IADIS’06), pages 406–410, San Sebastian, Spain, payment schemas. CASBA employs Java, JavaScript, April 2006. IADIS Press. [7] Grosof B. and Labrou Y. An Approach to using XML and a Rule-based CORBA and XML technologies, while advertising is e- Content Language with an Agent Communication Language. In Issues mail based. XML eases matching the data structures of in Agent Communication, number 1916 in Lecture Notes in Computer the CASBA ontology with those of the client database. Science, pages 96–117. Springer-Verlag, Heildeber, Germany, 2000. [8] Guttman R.H., Moukas A.G. and Maes P. Agent-Mediated Electronic • In [17] a model representing ontologies in a B2C scenario Commerce: A Survey. The Knowledge Engineering Review, 13(2):147– is proposed and a multi-agent architecture based on such 159, 1998. model is described. It realizes a virtual marketplace agent- [9] He M., Jennings N. R. and Leung H. On Agent-mediated Electronic Commerce. IEEE Transaction on Knowledge and Data Engineering, based where customers and merchants are supported by 15(4):985–1003, 2003. exploiting a representation of the concepts and behaviour [10] The Java Agent DEvelopment (JADE) framework website, 2006. involved in their EC world. Here, an agency has a central [11] Jennings N.R. and Wooldrige M.J. Applying Agent Technology. Int. Journal of Applied Artificial Intelligence, 9(4):351–361, 1995. role as mediator in coalition formation, agent communi- [12] Liu J. and Ye Y. Introduction to E-Commerce Agents: Marketplace cations and in virtual auctions. No payment scheme is Solutions, Security Issues, and Supply and Demand. In Proc. of the E- supported. Commerce Agents: Marketplace Solutions, Security Issues and Supply and Demand, volume 2033 of Lecture Notes in Computer Science, pages This aforementioned work exploit multi-agent systems to 1–6. Springer-Verlag, London, UK, 2003. support B2C activities. They are explicitly CBB based or [13] The North America Industry Classifications (NAICS) website, 2006. partially consistent with it. Customer interests and behaviour [14] O’Mahony D., Pierce M. and Tewari H. Electronic Payment Systems for E-Commerce. Artech House, Norwood, MA, 2nd edition, 2001. are taken into account in [17] where, similarly to CASBA, [15] Palopoli L., Rosaci D. and Ursino D. Agent’s roles in B2C E-commerce. authors exploit XML to realize a unified representation in AI Communications, 1912:95–126, 2006. order to reduce the impact of heterogeneities. Payment issues [16] Pinheiro R. Preventing Identity Theft Using Trusted Authenticatorss. Journal of Economic Crime Management, 2(1):1–16, 2004. are handled in MAGMA and in CASBA differently from [17] Rosaci D. A model of agent ontologies for B2C E-Commerce. In Proc. MAST. Finally, MAGMA is designed to support also hetero- of the Sixth International Conference on Enterprise Information Systems geneous agents whereas CASBA, [17] and MAST deal only (ICEIS’04), pages 3–9, Porto, Portugal, 2004. [18] Shamir A. SecureClick: A Web Payment System with Disposable Credit with homogeneous agents. Card Numbers. In Proc. of the 5th International Conference on Financial Cryptography, volume 2339 of Lecture Notes in Computer Science, VII. C ONCLUSION pages 232–242. Springer-Verlag, London, UK, 2002. [19] Sierra C. and Dignum F. Agent-mediated Electronic Commerce. Scien- This paper describes MAST, an XML-based multi-agent tific and Technological Roadmap. In Agent Mediated Electronic Com- system to support customers and merchants in a suitable, merce. The European AgentLink Perspective, volume 1991 of Lecture homogeneous and personalized way, taking into account their Notes in Artificial Intelligence, pages 1–18. Springer, 2001. [20] Tsvetovatyy M.B. and Gini M. Toward a Virtual Marketplace: Architec- interests on the basis of the behaviours shown during their tures and Strategies. In Proc. of the First International Conference on the B2C activities, represented as in CBB model. Furthermore, Practical Application of Intelligent Agents and Multi-Agent Technology the opportunities offered by XML (for agent profiles, the (PAAM’96), pages 597–613, London, United Kingdom, April 1996. [21] Vetter M. and Pitsch S. Towards a Flexible Trading Process over messages, the inter-catalogue representation of products and the Internet. In Agent Mediated Electronic Commerce, The European categories and the agent communication language) are used AgentLink Perspective, volume 1991 of Lecture Notes in Computer along with those of a secure centralized payment scheme, Science, pages 148–162. Springer-Verlag, London, UK, 2001. [22] W3C. Extensible Markup Language (XML) 1.1, February 2004. based on existing financial institutions and single-use account [23] Ye Y., Liu J. and Moukas A.G. Agents in Electronic Commerce - numbers (payments happen only among financial institutions, Special Issue on Agents in Electronic Commerce. Electronic Commerce over reserved communication channels, by preserving financial Research, 1(1-2):9–14, 2001. anonymity and confidentiality and benefiting of an existing authentication mechanism). Some results of experimental sim- ulations in a small B2C scenario, carried out using a Jade- based prototypal implementation of MAST, are presented. 82