A Novel Proposed Approach For Trust Calculation

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)
Web Site: www.ijettcs.org Email: editor@ijettcs.org

Volume 4, Issue 1, January-February 2015
ISSN 2278-6856
A Novel Proposed Approach for Trust

Calculation
Atul Saurabh1 and Swapnil Parikh2
1
BITS Edu Campus, Babaria Institute of Technology,

Varnama, Vadodara, Gujrat, India
BITS Edu Campus, Babaria Institute of Technology,

Varnama, Vadodara, Gujrat, India
Abstract
The world of internet provides many complex system and
services. Many of these systems provide good quality of
services which ends up in better user satisfaction. Other
services/systems do not satisfy user's need. Some of the
systems/services maintain user confidentiality and other
simply violate or provides low quality of confidentiality. In
this scenario, how a user can pick a service from service pool
and how to rely on that is a big challenging task. Many
methods are proposed for giving a numerical value for the so
called goodness provided by the system/services. This
numerical value gives a measure of system's trustworthiness.
The ranking is one of the methods. These methods mainly rely
on user's feedback and there is no role of measuring entity
other than collecting the feedback. Feedback is a strong
mechanism but suffering from a serious illness. Whether a
user giving feedback is lying or not is a big issue. Simply
working on the number of feedback is not enough for
calculating
the
trustworthiness.
Also
calculating
trustworthiness without concerning the system/service in
consideration is also not fair. In this paper we are proposing a
fair methodology for calculating the trust value of
system/service where the system/service will also be given a
chance for claiming. The calculated value is not biased
towards the user only. Rather it gives weightage to each
feedback received.
Keywords: Trust, Multi Agent System
1. INTRODUCTION
Today's world is internet world and more and more people
are getting connected. They share, store and retrieve vast
amount of information through numerous electronic
services provided on internet. It is a common belief that
whatever information is shared through these services
must be kept confidential. But fraudulent cases are also
not uncommon. Many technologies are employed to
overcome this problem. But technologies alone are not
enough. Trust is one of the key factor for electronic
services over internet. From the service providers' point of
view, how they can convince users' to trust their services.
Also from the users' point of view, how to trust systems
which provide reliable electronic services. Trust is actually
a belief in another. Trust is a subjective probability which
varies from 1 (complete trust) to 0 (complete distrust).
A related concept to the trust is reputation. In some of the
contexts trust is considered same as reputation but there
are some differences between them. Trust can be
Volume 4, Issue 1, January February 2015
calculated based on reputation. Reputation is the opinion

of an entity, whereas trust is derivation of reputation of an
entity.
Trust can be advantageous by number of ways:
1. Trust
can
eliminate much
of unnecessary
communication that may be required. It will improve
the performance.
2. Based on trustworthiness, decisions can be taken faster
and easier.
3. Trust is a kind of soft security compared to hard security
like encryption.
Trust can be estimated through number of ways:
1. Direct Experience: It is based on direct experience.
2. Communicated Experience: It is based on repeated
communication that has been done till now between
service provider and user.
3. Social Information: It is purely based on social
information.
4. Reputation: It is based on position of trustee in the
society.
2OVERVIEW
OF
PROBABILITY
AND
PROBABILISTIC MODEL
Set your page as A4, width 210, height 297 and
margins as follows:
2.1 Probability
Probability is a branch of mathematics that deals with
calculating the likelihood of a given event's occurrence,
which is expressed as a number between 0 and 1.
Mathematically probability can be expressed as:
(1)
Properties of probability
1. The sum of probabilities of an event and its
complementary is 1.
P (A) + P (Ar) = 1
(2)
2. The probability of impossible event is zero.
P () = 0
(3)
3. The probability of union of two events is the sum of
their probabilities minus their interaction.
P (A B) = P (A) + P (B) P (A B) (4)
4. If an event is subset of another events then its
probability is less than or equal to it.
If A < B then P (A) P (B)
(5)
Page 47

ISSN 2278-6856
P (D) = P (D | T ) P (T ) + P (D | T r) P (T r)
(7)
Where, T being the event that the theory is false.
many steps into the process. The Markova chain can be

demonstrated using a transition diagram. The state of the
diagram represents the intermediate point of the stochastic
process. The advancement from the one state to another
depends only upon the current state and not on the
previous state of the system. The above transition diagram
represents an evolution of the system where S1, S2, and
S3 are states of the system. The arrow represents the
transition from one state to another state. The weight of
the transition arrow represents the chance or the
probability for the transition.
2.3 Random Variable
If the system's current configuration is
2.2 Bayes' Theorem

Reverend Thomas Bayes' proved most important theory in
statistics: Let T denotes theory and D denote Data.
Then probability of theorem being true, given that the data
has been observed is
P (T | D) = P (D | T ) P (T )/P (D)
(6)
Where,
A Random Variable is a function or mapping f: E R

from event space to real number. In other words, a
random variable is a way to associate an event with a
number.
1. Let an experiment of tossing coin is made. The event
space is {H, T}. Let the function is defined as f(H)=3 and
F(T)=2. So, the event of coming up H is associated with 3
and that of T is associated with 2. So, f:ER is a random
variable.
2.4 Stochastic Process
A Stochastic Process (SP) is nothing but a collection of
random variables to express the evolution of any system.
The evolution of any system must have a start point and
final point. The start point is always known to us. But the
final point depends upon various conditions that a system
met during its evolution. Again, the system does not reach
final point immediately after start point. There may have
been many intermediate points. By introducing
intermediate points, a system is allowed to take any
random path to reach the final point. Also there may be
more than one final point in which system can stay.
The Stochastic Process normally depicted as in figure-1.
Figure 1 Stochastic Process

2.5 Markov Chain
As described earlier, Markov Chain is a stochastic process
in which future point (either intermediate point or final
point) depends only on the current intermediate point.
Markov property can also be called as memory less
property as the state of the system at future time tn+1 is
decided by the system state at the current time tn and does
not depend on the state at earlier time instance t1, t2, t3...,
tn. In general term, the distribution of where to go next
depends on where the system is now not on where the
system have been. Markov chains are the combinations of
probabilities and matrix operations. Markov chain models
a process that proceeds in steps (time, sequence, trial,
etc.); like a series of probability trees. This model can be
in one state in each step. When the next step occurs, the
process can be in the same state or move to another state.
These movements are defined by probabilities. One can
always find the probabilities of being in any given state
C=
then that means the probability that the system is in state
S1 is a, in S2 is band obviously in S3 is (1-(a+b)). Again if
probability that that the system can migrate from S1 to S2
is p1, S2 to S1 is p2 and so on then that matrix, called
transition matrix, can be represented as:
T=
We can predict the future state of the system from
equation-8
F=C*T
(8)
This matrix multiplication represents the next
configuration of the state. The row of resultant column
matrix represents the next probability of the
corresponding state.
3 LITERATURE SURVEY AND RELATED WORK
3.1 On How Agents Make Friends: Mechanism for
Trust Acquisition
Babak Esfandairi and Sanjay Chandrashekharan had
referred simple mechanisms for trust acquisition and
propagation. Authors had studied one to one trust
acquisition mechanisms. Authors had given propagation
model which is uses a directed graph methods to calculate
trust for human agents. A problem with their work is that
it does not make a distinction between distrust and lack of
knowledge about trust.
3.2 Bayesian Network-Based Trust Model
Yao Wang and Julita Vassileva proposed Bayesian
network based trust model. Authors had considered
situation where frequent interactions are there between
two agents. They have presented a flexible method for
presenting differentiated trust.
3.3 FIRE - An Integrated Trust and Reputation
Model for Open Multi-Agent Systems
T. Dong Huynh and Nicholas R. Jennings and Nigel R.
Shadbolt had presented novel model named FIRE which is
an integrated trust and reputation model for open multiagent systems. Authors claimed that FIRE can be easily
adapted to multi domains because of its modularized
design and parameterized configuration.
3.4 Modeling Trust in Multi-Agent Systems
Eli Stickgold, Sam Mahoney, Jonathan Pfautz , Joseph
Campolongo and Erik Thomsen had proposed graph based
Page 48

ISSN 2278-6856
approach for estimating trust for each agent in multi agent
systems. In order to achieve above task authors have used
Katz Centrality Matric as a measure of trust. The authors
claimed that the developed algorithm does not require any
hard facts and calculates trust based on relative
information.
3.5 SecuredTrust: A Dynamic Trust Computation
Model for Secured Communication in Multi-Agent
Systems
Anupam Das and M. Mahfuzul Islam presented a novel
and dynamic trust computation model called SecuredTrust
for evaluating agents in multi-agent environments. Here
authors had analyzed parameters related to the evaluation
of trust and then proposed comprehensive quantitative
model for calculating trust. They also proposed load
balancing algorithm based on analyzed parameters.
3.6 Trust Decision Making in Multi Agent System
Chris Burnett, Timothy J Norman and Katia Sycara
proposed their system considering risk, uncertainty and
high dynamicity. Authors had found a new approach to
select trustworthy partner. They have given five delegation
strategies for trust evaluation which are Simple
Delegation, Delegation with Monitoring, Delegate without
Monitoring, Delegate with Reputational Incentive and
Abstain from Delegation.
4 PROPOSED APPROACH
After going through some research papers, we are
proposing a new general approach for calculating trust for
any particular system. Figure-2 shows the flowchart for
proposed system. The proposed system has following steps
4.1 Calculating the Initial Trust Value
Initially the initiator will assign some work to the system
with supplied weight and will also calculate the chance of
failure. Like this, he will setup the initial trust value for
each and every system (the testing stage). Let us suppose
we have three suppliers S1, S2 and S3.
4.2 Bayesian Network-Based Trust Model
Now if at any point of time system S3 say complains that
the work get fail due to some other reason like delivery
system, then we will use Bayes theorem to calculate the
weight to truth. And the trust will be increased or
decreased accordingly.
4.3 Future Calculation
Now if a buyer is using any system say S2 then the
initiator can give the probabilistic view of the chance of
failure using the Markov Chain as we have the previous
data on which we can predict the future. Now if the buyer
reports the failure result, we again calculate the truth
weight of the buyer using Bayes Theorem and increase or
decrease the trust factor accordingly.
4.4 Future Calculation
If any system reports that it is recovered from the previous
illness, the initiator will first calculate Probability of Truth
using Markov Chain. If it is above the threshold value, the
initiator will put the system in testing stage. Otherwise the
trust will go down to zero.
Figure 2 Flow chart
5 PROPOSED APPROACH
5.1 Conclusion
We propose a new approach for calculating trust of a
particular system. Our proposed approach is parameter
independent as it uses weight of a parameter instead of
parameter, so it makes our system dynamic. Typical
mathematics used in proposed approach is: Probability
approach, Bayes theorem and Markov chain model.
5.2 Future Work
Future work of this report is to implement Trust function,
implementation of Markov chain model, implementation
of Bayes theorem, and implementation of proposed
approach. In the implementation of proposed approach,
number of parameter and weight of those parameters will
be taken as input and trust value is generated as output of
the approach.
References
[1] http://en.wikipedia.org/wiki/Multi-agent_system, last
access on 27th October, 2013.
[2] https://www.google.co.in/url?sa=t&rct=j&q=&esrc=s
&source=web&cd=1&cad=rja&ved=0CCYQFjAA&u
rl=http%3A%2F%2Fwww.lsi.upc.edu%2F~ia%2Fml
uck%2F1_rm.ppt&ei=ZcJsUvaTNZHirAe_xIH4CA&
usg=AFQjCNFEacH2dSQv5TZrPASwmId1YzMNKg
&sig2=a2JnlGJAOr0grxstpWYpRg&bvm=bv.551231
15,d.bmk, last access on 27th October, 2013.
[3] http://en.wikipedia.org/wiki/Procedural_reasoning_sy
stem, last access on 1st December, 2013.
Page 49

ISSN 2278-6856
[4] http://www.slideshare.net/easss2012/trust-andreputation-in-multiagent-systems, last access on 27th
October, 2013.
[5] K5. Babak Esfandiari, Sanjay Chandrasekharan, On
How Agents Make Friends: Mechanisms for Trust
Acquisition, Department of Systems and Computer
Engineering, Carleton University, Ottawa, Ontario,
Canada.
[6] Yao Wang, Julita Vassileva, Bayesian network-based
trust model, in Proceeding of IEEE/WIC
International Conference on WebIntelligence (WI),
Halifax, Canada, October 2003, pp. 372378.
[7] T. Dong Huynh and Nicholas R. Jennings and Nigel
R. Shadbolt, FIRE: An Integrated Trust and
Reputation Model for Open Multi-Agent Systems in
Proceedings of the 7th International Workshop on
Trust an Agent Societies, 2004, pp 65-74.
[8] Eli Stickgold, Sam Mahoney, Jonathan Pfautz, Joseph
Campolongo, Erik Thomsen, Modeling trust in
Multi Agent System, in Preceeding of the 22nd
Annual Conference on Behavior Representation in
Modeling and Simulation, Ottawa, Canada: BRIMS
Society, 2013.
[9] Anupam Das and M. Mahfuzul Islam, Secured
Trust: A Dynamic Trust Computation Model for
Secured Communication in Multi Agent Systems
IEEE Transactions on Dependable and Secure
Computing vol.9, no.2, year 2012.
[10] Chris Burnett, Timothy J. Norman, Katia Sycara,
Trust Decision-Making in Multi-Agent Systems,
Proceedings of the Twenty-Second International Joint
Conference on Artificial Intelligence , 2011.
presented and published 3 research papers in national and

international conferences. Also he has been appointed as a
reviewer in national and international conferences.
Currently he is working as department first year incharge. He is also working as a department coordinator for
Alumni Association and Workshops/Industry Visit.
AUTHOR
Atul Saurabh completed his B.Tech
from BITM Shantineketan (WB) in
2008. From Jan-2008-Nov- 2009 he
was involved in Web development in
Assurgent
Technology
Solution
Durgapur (WB) on JAVA/PHP domain. He has cleared
GATE-2010 with 91 percentile and took admission in
M.TECH which was completed in June 2012. During
M.Tech he was involved in various research activities on
Software engineering and also in the website development
for MRSI JUIT wings. Also he has achieved 2nd position
in an all India based competition on BIG Data organized
by PINGAR INDIA LTD, at Bangalore. In 2012 again he
cleared GATE-2012 with 91 percentile.
Prof. Swapnil M Parikh is working as
a Head (2nd Shift) and Assistant
Professor in Computer Science and
Engineering Department at Babaria
Institute of Technology, BITS edu
campus, Varnama since June 2007. He has obtained his
Masters' degree from Dharmsinh Desai University,
Nadiad. He has graduated in Information Technology
from Pune University, Pune. Prof. Swapnil. M. Parikh has
experience of 6.6 years in research and academia. He has
Page 50

A Novel Proposed Approach For Trust Calculation

Uploaded by

Copyright:

Available Formats

A Novel Proposed Approach For Trust Calculation

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Novel Proposed Approach For Trust Calculation

Uploaded by

Copyright:

Available Formats

International Journal of Emerging Trends & Technology in Computer Science (IJETTCS)

Web Site: www.ijettcs.org Email: editor@ijettcs.org