203

Document Sample
203 Powered By Docstoc
					   INTELLIGENT NEGOTIATION AGENT SYSTEM FOR B2C
                   E-COMMERCE
                                    Wen-Yau Liang,
    Department of Information Management, National Changhua University of Education,
                No.1, Jin-De Road, Changhua City 500, Taiwan, ROC
                              wyliang@cc.ncue.edu.tw


                                      ABSTRACT
      B2C e-commerce is becoming more widespread as more people come to
recognize its convenience and its ability to offer a quick response to requests and as
more products/services become available. However, many electronic marketplaces,
especially in the business-to-consumer, are in essence some kind of search engine
where buyers look for the best product in a database of products offered by sellers.
Usually such e-marketplaces do not use agent technology at all although agents could
significantly improve the services provided both for the buyers and the sellers.
Furthermore, negotiation capabilities are essential for B2C e-commerce systems. In an
automated negotiation, intelligent agents engage in broadly similar processes to
achieve the same end. In more detail, the agents prepare bids for and evaluate offers
on behalf of the parties they represent with the aim of obtaining the maximum benefit
for their users. Nevertheless, in current situation, price is the only criterion on which
agents are created. This factor is easy to measure and automate. However, criteria for
advanced transactions need to be more elaborating, e.g. giveback, dividend. In this
paper, we present a multiple-attributes negotiation model for B2C e-commerce, which
deploys intelligent agents to facilitate autonomous and automatic online buying and
selling by intelligent agents while providing fast response to consumers. These
include a 4-phase model, information collecting, searching/offer gathering,
negotiating, and evaluating. We also apply fuzzy theory and analytical hierarchy
process to develop the system interface to facilitate the user inputs. Finally, an
example of notebook buying process is illustrated.

Keywords: Intelligent agents, Multiple-attributes, Negotiation, B2C e-commerce

                                 1. INTRODUCTION
The internet has changed the environment of business fundamentally, because it offers
sellers and buyers a powerful communication channel and making it possible for the
two parties to come together in the e-marketplaces. B2C e-commerce is becoming
more widespread as more people come to recognize its convenience and its ability to
offer a quick response to requests and as more products/services become available.
Electronic commerce is a business practice associated with the buying and selling of
information, products, and services on Internet. Electronic commerce is increasingly
popular in today’s businesses (Du et al., 2005). There are more and more
organizations have paid attention to the advantage of this new tendency. So
E-commerce is now acting as a more and more important role in our daily life
(Anumba & Ruikar, 2002). Business-to-Consumer is similar in concept to the
traditional method of retailing, the main difference being the medium used to carry
out business by the internet. Such a method of carrying out business transactions
assumes that the consumer has access to the WWW. By selling direct to customers or
reducing the number of intermediaries, companies can achieve higher profits while
charging lower prices (Laudon & Laudon, 2000).
Negotiation is an inseparable component of many ecommerce activities, such as
auctions, scheduling, contracting, and so on, and is one area that can greatly benefit
from automation (Serguievskaia et al., 2006). A Negotiation Support System (NSS)
refers to a specialized group support system designed to help negotiators achieve
optimal settlements (Zhu et al., 2005). Software agent technology is a new approach
in e-negotiations. Use of software agents representing negotiating parties could
greatly decrease efforts and time needed to complete negotiations. Intelligent agent
software is the action of human decision-making behavior in the form of a computer
program. The intelligent agent software is that can help user to do some actions which
contain search, negotiation, trade off and so on to improve effectively. It also
improves the consumer’s bargaining position with the opposition by the internet and
traditional channels. In this paper, a multiple-attributes negotiation model for B2C
e-commerce is proposed.

           2. A SOLUTION APPROACH OF INTELLIGENT AGENT TO
                                    NEGOTIATION
Sycara et al. have precise definition that intelligent software agents are programs that
act on behalf of their human users to perform laborious information-gathering tasks
(Sycara et al., 2003). Other Scholars consider that agent architecture linking aspects of
perception, interpretation of natural language, learning and decision-making is
provided (Schleiffer, 2005). For B2C e-commerce applications, many varieties of
choice to the consumers have also introduced the problem of information overloading.
Meanwhile, there are so many e-shops and products for the consumers that it has
become too time-consuming to find the best deal (Wang et al., 2004). Thus, there is a
need for IA to assist in negotiation process for B2C E-commerce. The main purposes
of this paper are to develop a multiple-attributes negotiation model for B2C
e-commerce and provide more benefit and quickly response.
In this section, an agent-based architecture called an intelligent negotiation agent
(INA) architecture is designed to enhance the existing B2C e-commerce process
rather than to modify it, although the process may be modified before such a system is
built. The INA researches both the technology and the methods that are needed to
improve the way information is gathered, managed, distributed and utilized to
decision-makers in key business functions and operations. Several researches have
been studied on agent architecture. Nguyen and Jennings proposed model for the
buyer agent consists of three main components: a coordinator, a number of
negotiation threads and a commitment manager (Nguyen & Jennings, 2005). And
some researchers claimed that three key processes are needed in order to make agent
work effectively and simulate the real world buying experience: Identification of a
proper set of criteria on which to transact, Identification of agents with whom to
transact, Negotiation (Sarkis & Sundarraj, 2002).
In this paper, INA architecture which includes buyer agent and seller agent is
proposed. Buyer agent can search products, negotiate and access negotiation records.
The seller agent negotiates with buyer agent and access products and consumers
database. The architecture characteristics:
    Intelligent: The agent automatically customizes itself to the preferences of its
     customer (or client), based on previous experience and imprecise information
     from interaction with customers. The agent also automatically adapts to changes
     in its environment.
   Autonomous: An agent is able to take the initiative and exercise a non-trivial
    degree of control over its own actions through service agreements.
   Cooperation: An agent does not blindly obey commands, but makes suggestions
    to modify requests or ask clarification questions. It also cooperates with other
    agents to query the modules needed.
In the INA system, each INA is able to perform one or more services. A service
corresponds to some problem solving activities of negotiation. Service requirements
are issued either from other department, e.g., purchase department through an Intranet,
or from external customers through the Internet. Services are associated with one or
more agents that are responsible for the management and execution of those services.
Each service is managed by one agent, although the execution of it's sub-services may
involve a number of other agents. The activities of the INA agents involve:
    Selecting products to satisfy the requirements of customers
    Evaluating and Negotiation the products into an integrated service
    Coordinating and scheduling the processes intelligently.
All INAs have the same basic architecture. This involves an agent body that is
responsible for managing the agent's activities and interacting with peers and an
agency that represents the solution resources for the problems of product negotiation
processes. The body has a number of functional components responsible for each of
it's main activities – e.g. In buyer agent, interfacing with users, searching desired
products, negotiating with sellers and managing the tasks; In seller agent, interfacing
with users, negotiating with buyers and managing the tasks.

                             3. NEGOTIATION MODEL
According to the negotiation structure and flow as discuss above, we develop the
following negotiation model. The negotiation model includes negotiation decision
function (utility function), fuzzy theory and Analytical Hierarchy Process to get the
product utility. After that, applies the product utility to negotiate for the following
purposes:
    Decreasing the filtering time of product information.
    Improving the negotiation efficiency
    Satisfying buyer’s preference and maximum utility.

3.1 Utility Function
There are many papers which discuss agent negotiation and presented different
methods. Kraus et al. presented the issue of negotiation time (Kraus et al., 1995).
Additionally, Faratin et al. presented the Negotiation Decision Function (NDF) which
was the negotiation criterion (Faratin et al., 1998). This method allows agents
negotiate with multi-attributes such as price and quantity. The negotiation strategy
includes time, resource and so on. In this paper, NDF function is extended to decrease
negotiation times.

3.2 Analytic Hierarchy Process
The Analytic Hierarchy Process (AHP) is a mathematical decision making technique
that allows consideration of both qualitative and quantitative aspects of decisions. The
AHP method uses the human ability to compare single properties of alternatives. It not
only helps decision makers choose the best alternative, but also provides a clear
rationale for the choice. The process was developed in the 1980 by Saaty (Saaty,
1980). In recent years, the AHP has already been studied extensively. Gu and Zhu
make the decision method of many attribute according to the vector quantity of the
characteristic (Gu & Zhu, 2006). Osman and Cengiz also propose Fuzzy
multi-attribute selection among transportation companies using axiomatic design and
analytic hierarchy process (Osman & Cengiz, 2005).

3.3 Fuzzy Theory
Fuzzy logic is a superset of conventional logic that has been extended to handle the
concept of partial truth values between "completely true" and "completely false". It
was introduced by Zadeh in the 1965 as a means to model the uncertainty of natural
language (Zadeh, 1965). The fuzzy theory has already a lot of researches at present.
For example, Fashiem et al. use a fuzzy logic-based system for assessing the level of
business-to-consumer (B2C) trust in electronic commerce (Fahim et al., 2005). Cheng
et al. apply the intelligent agent into the electronic market and the trade negotiation is
by the fuzzy inference system (Cheng et al., 2006). In this paper, we will apply the
theory of fuzzy set and define attributes membership function to get attributes margin
utility value.

3.4 Negotiation Strategy
In this paper, both buyer agent and seller agent own their negotiation strategy. Buyer
strategy means the offer method of buyer agent and the stop conditions. This offering
function is defined as follow:
   Offernew  utility  100  u  Offerold


Which Offer means new price, utility is product utility, u is the unit increase value,
           new


Offer
      is the last offer. Beside current offer, buyer agent must know when to stop
       old

negotiate. In this paper, we present two conditions which both of them must be
reached and then agents can decide trade or not. The first condition is that the product
price which seller agent present must in buyer offer range. The succeed condition is
that the ratio of buyer offer and seller offer must larger than a threshold which defined
by buyer in negotiation initial stage. Seller strategy decides the seller agent current
offer and the stop condition. This paper calculate next offer by last offer as follow
(Fernando et al., 2000).
    x[i] new  x[i] old  (1) w F | RVi  x[i] old |



x[i ]new     is the new offer and   x[i ]old   is the last offer. F is the factor which between 0 to 1,
w is factor to control increase or decrease. RV means the max or min limit value,
setting value or buyer offer.

3.5 Negotiation Process
The negotiation process can divide four stages which include information collecting,
searching/offer gathering, negotiating, and evaluating. Figure 2 shows the negotiation
processes workflow. The first stage includes insert product search and negotiation
conditions, setting product attribute membership function, compare product attribute
and apply AHP to calculate attribute weights. Second stage is according to search
conditions that user inserts. And then agent will search products from internet and get
sellers’ information response to user. Thirdly, agents will start negotiate by the search
result and calculate product utility and determine whether receive product or not.
Finally, negotiation evaluate state is the finally state of negotiation which buyer agent
and seller agent will determine when to finish.

                            4. EXAMPLE ILLUSTRATION
In this paper, an example of instance notebook sales in an open e-commerce platform
is illustrated. The sellers will provide the information of the products on e-commerce
platform through the seller agent. The buyer also will filtrate, search and negotiate
with the seller through the buyer agent. The four negotiation stages which presented
above will be described and show below.

4.1 Information Collecting Stage
Before negotiating, buyer agent will ask user to login the system and input preference,
related settings which include price ratio, price range, product specifications,
giveaway and preference. Then, the membership function of the products setting is let
user to set all products attribute in order to carry on the follow-up negotiation by the
agent. The system will then ask user to compare the product attribute. The system will
calculate the product attribute weight by the AHP.

4.2 Searching/Offer Gathering Stage
After buyer input the search conditions and define membership functions, agent will
go to next stage and search products from e-commerce platform.

4.3 Negotiating Stage
In this step, according to search result, buyer agent will calculate product utility. We
divide this stage into two steps. Firstly, buyer agent will calculate attribute utility.
Secondly, apply utility function and attribute weight to get product utility. In this
system, it will show the negotiation processes that contain buyer's offer and seller's
offer. Finally, when negotiation ends, it will show products and price that both sides
can accept.

4.4 Evaluating Stage
When buyer agent receives offer from seller agent, buyer agent will compare the offer
with buyer price range first. The price which seller agent presents must lower than
upper limit. After that, buyer agent also evaluates the product utility which must larger
than utility threshold. On the other hand, if buyer agent does not receive seller agent
prices, then buyer agent will calculate new prices and present to seller agent. After
receive buyer agent prices, seller agent will check the prices which must larger than
the lower limit of price range.

                                 5. CONCLUSION
This paper we present negotiation model which include utility function, fuzzy theory
and AHP in B2C e-commerce environment. Agents support both buyers and sellers to
negotiate each other and then present benefit and response quickly. In future work, we
can validate the model and develop the negotiation system. After that, analysis the
feasibility on actually world and then modify the model to more fit the B2C
e-commerce environment.

                              ACKNOWLEDGEMENT
This work was partially supported by funding from the Nation Science Council of the
Republic of China (NSC 95-2416-H-018-010).

                                  REFERENCES
Anumba, C.J. & Ruikar, K. 2002. Electronic commerce in construction—trends and
    prospects. Automation in Construction, 11(3), 265-275.
Cheng, C.B., Chan, C.C.H., & Lin, K.C. 2006. Intelligent agents for e-marketplace:
    Negotiation with issue trade-offs by fuzzy inference systems. Decision Support
    Systems, 42(2), 626-638.
Du, T.C., Li, E.Y., Chou, D. 2005. Dynamic vehicle routing for online B2C delivery.
    Omega, 3(1), 33-45.
Fahim, A., Dave, H. & Zakaria, M. 2005. A fuzzy logic-based system for assessing the
    level of business-to-consumer (B2C) trust in electronic commerce. Expert
    Systems With Applications, 28(4), 623-628.
Faratin, P., Sierra, C. & Jennings N.R. 1998. Negotiation Decision Functions for
     Autonomous Agents. Robotics and Autonomous Systems, 24, 159-182.
Fernando, L., Nuno, M., Novais, A.Q. & Helder, C. 2000. Towards a generic
     negotiation model for intentional agents. International Workshop on Database
     and Expert Systems Applications, 433-439.
Gu, X. and Zhu, Q. 2006. Fuzzy multi-attribute decision-making method based on
     eigenvector of fuzzy attribute evaluation space. Decision Support Systems, 41,
     400-410.
Kraus, S., Wilkenfeld, J., & Zlotkin, G. 1995. Multiagent negotiation under time
     constraints. Artificial Intelligence, 75(2), 297-345.
Laudon, K.C. & Laudon, J.P. 2000. Management Information Systems : Organization
     and technology in the networked enterprise, 6th edn, Prentice-Hall, London.
Nguyen, T.D. & Jennings, N.R. 2005. Managing commitments in multiple concurrent
     negotiations. Electronic Commerce Research and Applications, 4(4), 362-376.
Osman, K. & Cengiz, K. 2005. Fuzzy multi-attribute selection among transportation
     companies using axiomatic design and analytic hierarchy process. Information
     Sciences, 170, 191-210.
Saaty, T.L. 1980. The Analytic Hierarchy Process. New York: McGraw Hill.
Sarkis, J. & Sundarraj, R.P. 2002. Evolution of brokering paradigms in e-commerce
     enabled manufacturing. International Journal of Production Economics, 75,
     21-31.
Schleiffer, R. 2005. An intelligent agent model. European Journal of Operational
     Research, 166(3), 666-693.
Serguievskaia, I., Al-Sakran, H. & Atoum, J.O. 2006. A Multi-Agent Experience
     Based e-Negotiation System. Information and Communication Technologies,
     2006. ICTTA '06. 2nd, 1(24-28), 286-291.
Sycara, K., Paolucci, M., Ankolekar, A. & Srinivasan, N. 2003. Automated discovery,
     interaction and composition of Semantic Web services. Web Semantics: Science,
     Services and Agents on the World Wide Web, 1(1), 27-46.
Wang, Y., Tan, K.L. & Ren, J. 2004. PumaMart: a parallel and autonomous agents
     based internet marketplace. Electronic Commerce Research and Applications,
     3(3), 294-310.
Zadeh, L.A. 1965. Fuzzy sets. Information and Control, 8, 338-353.
Zhu, F., Guan, J., Wang, Y., Zhou, J. & Liao, B. 2005. An automatic negotiation
     method based on CBR and agent reasoning. Computer and Information
Technology, 2005 ( CIT 2005 ) . The Fifth International Conference, pp.
1019-1023.

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:3
posted:12/7/2011
language:
pages:7