J. Software Engineering & Applications, 2009, 2: 103-110
doi:10.4236/jsea.2009.22015 Published Online July 2009 (www.SciRP.org/journal/jsea)
Autonomic Software Component QoS Matchmaking
Algorithm Based on Fidelity Factor in Agent-Based
Autonomic Computing System
Kun Zhang1,2, Manwu Xu2, Hong Zhang1
1School of Computer Science & Technology, Nanjing University of Science & Technology, Nanjing, China; 2State Key Laboratory
for Novel Software, Nanjing University, Nanjing, China.
Received December 8th, 2008; revised March 20th, 2009; accepted March 31st, 2009.
ABSTRACT
Autonomic software component (ASC) QoS matchmaking problem for autonomic element has been taken as one of the
most important issue in field of autonomic computing based on agent. Aimed at overcoming drawbacks such as subjec-
tiveness and unfairness, and improving the self-configuration capability for autonomic element, we introduce evalua-
tion mechanism of confidence of individual QoS attributes during ASC QoS matchmaking, i.e., fidelity factor for each
attribute, and propose an ASC QoS matchmaking algorithm based on fidelity factor. Simulation experiments demon-
strate that our proposed algorithm performs best performance in terms of QoS than other existing algorithms, and has
better compromise between attribute quality and users’ evaluation when selecting ASC.
Keywords: Autonomic Software Component, QoS Matchmaking, Fidelity Factor, Autonomic Computing, Autonomic
Element
1. Introduction
In mid-October 2001, aimed at the problem of looming
software complexity crisis, IBM Company innovatively
proposed autonomic computing [1] technology – com-
puting systems that can manage themselves given
high-level objectives from administrators. An autonomic
software component (ASC, or element, in IBM parlance)
[1] is—
“The fundamental atom of autonomic applications and
systems—a modular unit of composition with contractu-
ally specified interfaces, explicit context dependencies,
and mechanisms for self management, responsible for
providing its own services, constraints (e.g., system re-
source requirements, performance requirements, etc.),
managing its own behavior in accordance with context,
rules, and policies, and interacting with other autonomic
compo n ents. [1]”
Autonomic software components will have complex
life cycles, continually carrying on multiple threads of
activity, and continually sensing and responding to the
environment in which they are situated. Autonomy, pro-
activity, and goal-directed interactivity with their envi-
ronment are distinguishing characteristics of software
agents. Viewing au tonomic software components as a g en t s
and autonomic systems as multi-agent systems makes
it clear that agent-oriented architectural concepts will be
critically important [2].
In order to efficiently accomplish the self-configurati-
on and self-management between ASCs, one of the most
important issues is how to design an efficient autonomic
software component matchmaking algorithm to find ap-
propriate provider(s) for a consumer according to the
providers’ advertising description.
However, most of the existing algorithms are only
concerned with functional matchmaking on software
component. Few of them focused on non-functional fac-
tors, such as software component cost, time, reliability,
satisfaction, i. e. quality-of-service (Qo S) of software co-
mponent. Therefore, when more ASCs than one can pro-
vide a functional component, these algorithms just ran-
domly selected one from these ASCs, and could not be
obtained optimal ASC with high QoS performance.
Moreover, only a few existing algorithms are considered
QoS factors during software component matchmaking.
However, these algorithms were paid regard to QoS level
as a whole, ignoring confidence of individual QoS at-
tributes. Due to QoS data was often advertised by soft-
ware component providers, it suffered from the draw-
backs such as subjectiveness and unfairness. And these

Autonomic Software Component QoS Matchmaking Algorithm Based on Fidelity
B
104 Factor in Agent-ased Autonomic Computing System
algorithms didn’t be considered the user’s feeling and
satisfaction.
Aimed at overcoming the above difficulties, we intro-
duce evaluation mechanism of confidence of individual
QoS attributes during ASC QoS matchmaking, i.e., fidel-
ity factor for each attribute, and propose an ASC QoS
matchmaking algorithm based on fidelity factor, which
will improve the facility and fairness concerned with
QoS attributes, and objectively select optimal ASC with
QoS performanc e.
The rest of the paper is organized as follows. Some of
the existing work related to this paper is described in
Section 2. Section 3 gives our proposed ASC match-
making algorithm in detail. A case study and simulation
results are presented in Section 4. Section 5 concludes
the paper.
2. Related Work
In recent years, much research has been devoted to the
development of software component matchmaking algo-
rithms for agent or autonomic element [3,4,5,6,7,8].
Wickler [3] addressed the problem of capability broker-
ing agent, and proposed a new capability description
language (CDL) for the representation of agent capabili-
ties. Sycara [4] defined a language called LARKS for
agent advertisements and requests, and presented a flexi-
ble and efficient matchmaking process that used LARKS.
LARKS performed both syntactic and semantic matching,
and in addition allowed the specification of concepts
(local ontologies) via ITL, a concept language. The es-
tablishment for semantic distance consumed much
workload, so that their matching algorithm had limita-
tions on practicability and reliability. Arisha et al. [5]
provided approximate software agent service match-
making by using semantic distance. Whereas, the algo-
rithm could not support definition of data type and de-
script software component efficiently. However, neither
of the above algorithms considered the factors of qual-
ity-of-service attribution, such as cost, time, and reliabil-
ity. Zhang [6] and Jiang [7] analyzed the drawbacks in
previous papers, i.e., the matchmaking was only based
on the advertised capabilities of provider agents or soft-
ware components. They considered the matchmaking
influenced by QoS of software components, and indi-
vidually presented software components matchmaking
algorithm based on QoS. In [6], it was argued that the
practical performance of provider agents had a signifi-
cant impact on the matchmaking outcomes of middle
agents. The authors’ proposed algorithm could pick up
the provider agents based on the history information in
accomplishing similar tasks in the past rather than
choosing randomly. At th e launching of an agent system,
the proposed algorithm provided initial values of the
track records. With agents’ history information and the
initial values of the track records, the quality of match-
making algorithms could be improved significantly, and
the returned results were more accurate and reasonable.
Jiang [7] pointed out that there were two drawbacks us-
ing the track records in [6], i.e., the value of track re-
cords was too subjective, and track record model is too
simple to judge agent services’ performance more accu-
rately. So, the authors in [7] presented QoS-driven
matchmaking algorithm, which aimed at matching the
best satisfying agent for u ser. It is pity for the above two
algorithms that the authors didn’t consider the user’s
feeling and satisfaction when selecting agents or soft-
ware components. The authors in [8] studied agent ser-
vice selection and matchmaking in manufacturing indus-
try field. They applied the idea of rough sets theory and
fuzzy information filter to reclaim agent selection prob-
lem and realized reasonable evaluation for recycling
quotient. Their algorithm was easy to select optimum
reclaiming agent for manufacturer. However, their
method belonged to specific field and non-universal.
Besides software component or agent service match-
making, some researchers took software component or
agent as web service, and proposed many QoS-aware
web services matchmaking or selection algorithms
[9,10,11,12,13,14]. Some theoretic methods or principle
in these algorithms can be used for reference when
matchmaking software component, du e to many similari-
ties between software component and web service. Ma [9]
proposed a semantic QoS-aware framework for semantic
web services discovery, and presented a selection algo-
rithm to obtain the optimal offer from clients’ viewpoint
under complex QoS conditions through confirming the
compatibility of concepts. Hu et al. [10] proposed a
novel and extended web service QoS model (attributes
include time, co st, reliability and interest) by adding an
interest degree property, and then put forward to a QoS
matchmaking algorithm based on this model. Aimed at
resolving such conflicts to ensure consensus on the QoS
characteristics in the selection of web services, Wei [11]
proposed a QoS Consensus Moderation Approach
(QCMA) in order to perform QoS consensus and to alle-
viate the differences on QoS characteristics in the selec-
tion of web services. QCMA enhanced the moderation
for opinion similarity and preference on QoS attributes
and was been designed a mechanism for providing group
consensus moderation on QoS. Giallonardo et al. [12]
used ontology to describe QoS-based WS specifications
and reasoners to perform the matchmaking. They ad-
dressed QoS semantics to improve the recall of the
matching process by exploitin g ontology knowledg e, and
their QoS specification approach used metrics to under-
stand, describe, and control the QoS in the matching

Autonomic Software Component QoS Matchmaking Algorithm Based on Fidelity 105
Factor in Agent-Based Autonomic Computing System

process. However, they didn’t offer useful results for
over-constrained demands. In addition, although they
were capable of, they did not perform semantic QoS
metric matching. Guo et al. [13] proposed three-dimen-
sional QoS model of web services, and designed a web
services selection algorithm. In their algorithm, the con-
cept and measure method of web service effectiveness
were proposed, and th en a web service rank ing algorith m
based on the effectiveness of web service was designed.
The comprehensive experiment showed their proposed
model possessed high precision, high response rate and
better influence on the load balance of web service. Li-
ang Kai-jian [14] studied QoS support problem in Ser-
vice-oriented Grid system, and presented a new parame-
ter called GSQN(Grid Service Quality Number) to depict
the profile of qualitative characteristics o f a grid service.
Based on this idea, the author proposed a strategy of
matchmaking with a price model to match grid service.
3. Our Proposed Matchmaking Algorithm
3.1 Problem Description
Definition 1. Autonomic software component match-
making based on QoS problem. Given the set of ASC
providers , an ASC re-
quester rasc, autonomic software component match-
making based on QoS problem is to find an ASC such
that their QoS level is maximized.

12
,,..., n PASCpasc pascpasc
3.2 QoS Model for Autonomic Software
Component
Definition 2. Autonomic software component QoS mod-
el for autonomic element is a 6-tuple as follows:
_
_(,,,,,qosmodelasct c relm repfid) (1)
where t is ASC response time from sending request to
receiving result, including process time and transmission
delay time, i.e., qtime(asc) = Tprocess(asc) + Ttrans(asc); c is
ASC cost, and represents the expenses paid by users to
ASC provider; rel is a metric of reliability, denoting the
probability of ASC providing its registered software
component; rel is a technical measure related to hard-
ware and/or software configuration of ASCs and the net-
work connections between the ASC requesters and pro-
viders. m denotes the probability of accurate mainte-
nance when an exception occurs for an ASC; rep is a
measure of ASCs’ trustworthiness or satisfaction d egree,
and denotes users’ (or requesters’) satisfaction to ASC.
rep mainly depends on end user’s experiences of using
the ASC. Different end users may have different opin-
ions or satisfaction degree on the same ASC. Usually,
the end users or requesters are given a range to rank or
score ASCs, for example, in Amazon.com, the range is
[0,5]. In this paper, the range is [0,1 ].
The first four QoS attributes are often published by
ASC providers themselves, and describe the basic QoS
performance of an ASC, so most of the authors took
them as quality criteria for an ASC. However, such crite-
rion is too subjective to reflect actual quality of ASC. In
this paper, we introduce attribute “reputation” to meas-
ure users’ satisfaction degree. On the other hand, fidelity
vector is added to the QoS model to improve the impar-
tiality and objectivity. Fidelity is treated as a vector
composed of fidelity attributes. Each fidelity attribute
refers to the confidence or fidelity of above first four
QoS attributes, i.e.,
,, ,
tcrelm
f
idfid fidfidfid (2)
3.3 Matchmaking Algorithm Description
In our proposed QoS matchmaking algorithm, the draw-
backs in [6] and [7] such as subjectiveness and unfair-
ness are overcame through introducing evaluation
mechanism of confidence of individual QoS attributes,
i.e., fidelity factor for each attribute, which can improve
the self-configuration capability for autonomic element.
The basic idea of our algorithm is as follows: firstly,
normalizing each attribute in QoS description to range
[0,1] for each ASC in initial set; then fidelity of each
attribute is considered to evaluate QoS overall perform-
ance for every ASC comprehensively and objectively;
finally, an ASC whose total QoS value is maximal is
picke d out from all candidates.
Concretely, suppose that there is a set of ASC pro-
viding a certain software function, i.e.PASC



12
,,..., n pasc pascpasc. By merging the quality vector s
of all these candidates, a matrix


,ij
is built according with Definition 2, in which each row
Qj corresponds to an ASC pasci while each column
corresponds to a QoS attribute value. Let a matrix
,1QiQ,1 5nj 


4
,,1 ,1 ij
fidi nj

 Fdenote confidence fidelity
of the first four QoS attributes for pasci. There are two
phases as follows:
1) Scaling phase.
Some of the QoS attributes could be negative, i.e., the
higher the value, the lower the quality, such as time, cost.
Other QoS attributes are positive, i.e., the higher the
value, the higher the quality, such as reliability, main-
tainability, and reputation. For negative attributes, val-
ues are scaled according to (3). For positive criteria, val-
ues are scaled according to (4).
max ,max min
max min
,
max min
0,1,
10
jij
jj
jj ij
jj
QQ if QQ
QQ Mj
if QQ






2 (3)

Autonomic Software Component QoS Matchmaking Algorithm Based on Fidelity
Factor in Agent-Based Autonomic Computing System
106
min
,max min
max min
,
max min
,0
,3,4,
10
ij j
jj
jj ij
jj
QQ if QQ
QQ M
if QQ
 





5j (4)
In the above equations, and are the maxi-
mal and minimal value of the jth QoS attribute, respec-
tively. Let ,ij be normaliz-
ing matrix according to Q, where ,ij
max
j
Q
,1
min
j
Q

5j,1MinM
M
denotes nor-
malizing value of the jth QoS attribute for ASC provider
pasci.
Proof. In scaling phase, the time complexity of getting
normalizing matrix for all attributes is O(kn), where k is
the number of attributes (constant), in this work, k=5.
Then, weighting phase takes O(n). So, the worst time of
algorithm is O(n).
4. Simulation Results
4.1 A Case Study
A case study of our proposed matchmaking algorithm
will be given for explaining the effect of our proposed
algorithm. The following experiment method will be
used to select the most satisfactory ASC. Suppose the
initial providers have 10 ASCs, i.e.,PASC



12 1
,,,pasc pascpasc0 . It means that these ten ASCs
can provide the same or closely similar capabilities.
2) Weighting p hase.
The following formula is used to calculate the overall
quality score of ASC provider pasci.
4
,,5
1
()( )
ij ijiji
j
qos pascwMfidwM


,5
1
(5) The QoS values of 10 ASCs are gener ated through the
following simulation. Assume that there are 50 similar
tasks (or ASC requests). For each request, we randomly
delegate it to an ASC, pasci, from PASC, and randomly
generate QoS attribute values and their fidelity values,
denoting QoS value of pasci in this request. In these at-
tributes, t is randomly distributed between 70 and 100 ms;
c is uniformly distributed [10, 100] $; rel, m, rep is ran-
domly generated in [0,1]. The fidelity for the first four
attributes is in [0,1]. The QoS values of 10 ASCs for 50
requests are shown in Table 1, where pn denotes pro-
vided ASC nu mber of pasci for all requests.
where, is the weight value of the jth QoS attribute,
. End users express their prefer-
ences regarding QoS by providing values for the weight wj.
j
w
01w
5
1
,
jj
j
w



3.4 Algorithm Analysis
Theorem 1. In worse case, the time complexity of our
proposed algorithm is O(n), where n is the total number
of ASC providers.
Table 1. The QoS values of 10 ASCs for 50 request
pasc pn QoS attribute value. (time, cost, reliability, maintainability, reputation)
pasc1 4
(81.364, 94.46, 0.069, 0.614, 0.575), (73.451, 35.37, 0.838, 0.042, 0.187), (71.636, 99.96, 0.304, 0.106,
0.656), (99.294, 43.29, 0.384, 0.493, 0.011)
pasc2 5
(84.663, 23.28, 0.740, 0.456, 0.716), (79.636, 70.50, 0.746, 0.767, 0.616), (91.45, 26.392, 0.427, 0.235,
0.245), (91.586, 92.89, 0.124, 0.888, 0.587), (73.615, 87.73, 0.738, 0.082, 0.804)
pasc3 3
(99.547, 94.69, 0.690, 0.593, 0.930), (90.360, 86.49, 0.422, 0.235, 0.356), (90.825, 22.23, 0.678, 0.784,
0.763)
pasc4 4
(77.087, 90.03, 0.947, 0.393, 0.697), (74.888, 44.23, 0.978, 0.834, 0.624), (80.114, 12.45, 0.502, 0.948,
0.990), (73.533, 35.89, 0.114, 0.325, 0.026)
pasc5 9
(82.862, 32.69, 0.708, 0.043, 0.986), (73.595, 59.18, 0.611, 0.816, 0.219), (97.835, 69.63, 0.150, 0.017,
0.724), (91.110, 11.66, 0.540, 0.945, 0.221), (97.034, 54.32, 0.205, 0.832, 0.493), (93.738, 58.43, 0.659,
0.478, 0.182), (70.386, 36.56, 0.674, 0.919, 0.364), (70.894, 60.31, 0.288, 0.810, 0.637), (98.963, 65.05,
0.412, 0.013, 0.052)
pasc6 4
(73.856, 41.60, 0.411, 0.647, 0.385), (89.247, 25.84, 0.409, 0.542, 0.114), (87.707, 16.07, 0.649, 0.706,
0.112), (88.005, 90.24, 0.991, 0.102, 0.045)
pasc7 5
(91.256, 34.28, 0.566, 0.295, 0.087), (91.973, 61.65, 0.708, 0.659, 0.263), (89.541, 24.46, 0.096, 0.515,
0.247), (89.445, 27.47, 0.342, 0.484, 0.978), (85.628, 31.53, 0.692, 0.551, 0.706)
pasc8 8
(82.750, 60.07, 0.224, 0.790, 0.624), (97.355, 89.42, 0.283, 0.307, 0.567), (71.577, 88.84, 0.643, 0.443,
0.804), (80.630, 51.27, 0.360, 0.116, 0.666), (97.848, 46.98, 0.243, 0.958, 0.529), (82.601, 50.14, 0.681,
0.785, 0.296), (75.847, 14.88, 0.266, 0.178, 0.547), (87.174, 71.81, 0.983, 0.453, 0.635)
pasc9 4
(93.761, 60.72, 0.447, 0.177, 0.424), (75.342, 47.91, 0.730, 0.548, 0.783), (77.398, 70.89, 0.231, 0.295,
0.638), (78.867, 40.35, 0.624, 0.599, 0.457)
pasc10 4
(88.814, 56.19, 0.883, 0.563, 0.029), (80.738, 64.78, 0.728, 0.648, 0.872), (75.470, 20.82, 0.637, 0.401,
0.752), (72.877, 20.72, 0.305, 0.750, 0.601)

Autonomic Software Component QoS Matchmaking Algorithm Based on Fidelity 107
Factor in Agent-Based Autonomic Computing System






















The QoS values in matrix Q and the confidence fidel-
ity matrix of the first four QoS attributes for ASC pasci
are the mean value of the results produced by its pro-
vided software component number, respectively. The
QoS values matrix Q, confidence fidelity matrix F, and
normalizing matrix M are as follows.
81.44, 68.27 , 0.40, 0.31, 0.36
84.19, 60.16 , 0.56, 0.49, 0.59
93.58, 67.80, 0.60, 0.54, 0.68
76.41, 45.65, 0.64, 0.62, 0.58
86.27, 49.76 , 0.47, 0.54, 0.43
84.70, 43.44 , 0.62, 0.50, 0.16
89.57, 35.88, 0.48
Q
, 0.50, 0.46
84.47, 59.18 , 0.46, 0.50, 0.58
81.34, 54.97 , 0.51, 0.40, 0.58
79.47, 40.63, 0.64, 0.59, 0.56
















(6)
0.34, 0.15, 0.41, 0.40
0.68, 0.47, 0.81, 0.36
0.54, 0.53, 0.41, 0.62
0.41, 0.76, 0.38, 0.26
0.51, 0.69, 0.43, 0.65
0.35, 0.30, 0.46, 0.48
0.60, 0.18, 0.66, 0.67
0.51, 0.42, 0.37, 0.46
0.42, 0.65, 0.50, 0.47
F
0.51, 0.65, 0.45, 0.47
















(7)
0.71, 0.00, 0.00, 0.00, 0.37
0.55, 0.25, 0.65, 0.55, 0.83
0.00, 0.01, 0.83, 0.72, 1.00
1.00, 0.70, 0.99, 1.00, 0.81
0.43, 0.57, 0.31, 0.73, 0.51
0.52, 0.77, 0.90, 0.60, 0.00
0.23, 1.00, 0.34, 0.60, 0.56
0.
M
53, 0.28, 0.26, 0.61, 0.81
0.71, 0.41, 0.46, 0.29, 0.79
0.82, 0.85, 1.00, 0.89, 0.77
















(8)
Suppose that weight values of each QoS attribute de-
fined by users are: w1=0.15, w2=0.2, w3=0.2, w4=0.15
and w5=0.3. The overall quality scores of each ASC pro-
vider by using Formula (5) are shown in Table 2.
It is easy to see from Table 1 that pasc10 has maximal
QoS score among all ASCs, and this ASC will be se-
lected as optimal one with QoS performance.
The matchmaking result ASCs and their normalizing
matrix are shown in Table 3 by using traditional algo-
rithm, Zhang’s algorithm [6] and Jiang’s algorithm [7],
respectively.
In traditional matchmaking algorithm, the first ASC is
usually selected as match result, i.e., pasc1. In Zhang’s
algorithm, the track records concept is equivalent to the
fifth QoS attribute in our algorithm, i.e., reputation. So,
the result QoS scores of each agent are: -1, 1.33, 1.33,
0.67, -1.33, -3.33, -1, 1.33, 0.67, 1. As a result, pasc2, or
pasc3, or pasc8 with maximal value (1.33) will be se-
lected. Finally, Jiang’s algorithm will be compared with
our proposed algorithm, and fidelity factor is not consid-
ered by Jiang’s algorithm. The overall QoS scores of
each ASC by using their algorithm are: 0.218, 0.594,
0.576, 0.880, 0.503, 0.501, 0.562, 0.521, 0.562, 0.858.
Obviously, pasc4 (score is 0.880) is optimal ASC.
We can see from normalizing matrix M in Table 2,
QoS performance (0.82, 0.85, 1.00, 0.89, 0.77) of pasc10
gives slightly worse than performance (1.00, 0.70, 0.99,
1.00, 0.81) of pasc4. But, fidelity performance of QoS
attributes of pasc10 (0.51, 0.65, 0.45, 0.47) is evidently
better than that of pasc4 (0.41, 0.76, 0.38, 0.26). This
phenomenon shows that some ASC providers usually
claim their higher QoS attribute performance like pasc4,
but if the confidence fidelity or users’ feeling degree is
also considered when selecting ASC, the overall quality
or performance for these ASCs is not necessarily optimal.
This point accords with the practical situation on Internet,
such as e-business, online shopping, in which, people
have always compromised selection need software func-
tion between quality and users’ evaluation. Therefore,
our proposed QoS matchmaking algorithm is effective
and reasonable.
Table 2. The QoS score for each ASC provider
ASC overall QoS ASC overall QoS
1 0.147 6 0.200
2 0.463 7 0.332
3 0.437 8 0.368
4 0.525 9 0.402
5 0.363 10 0.558
Table 3. Results by using different algorithms
Algorithm ASCQoS normalizing value
Traditional Alg. 1 (0.71, 0.00, 0.00, 0.00, 0.37)
Zhang’s Alg. 2 3
8
(0.55, 0.25, 0.65, 0.55, 0.83)
(0.00, 0.01, 0.83, 0.72, 1.00)
(0.53, 0.28, 0.26, 0.61, 0.81)
Jiang’s Alg. 4 (1.00, 0.70, 0.99, 1.00, 0.81)
Our proposed Alg.10 (0.82, 0.85, 1.00, 0.89, 0.77)

Autonomic Software Component QoS Matchmaking Algorithm Based on Fidelity
B
108 Factor in Agent-ased Autonomic Computing System
4.2 Simulation Experiment
To evaluate the impact of fidelity on the final match-
making results, we conduct the following simulations
and compare the matchmaking results of our algorithm
with Jiang’s algorithm.
For the sake of simple, suppose that there are 10 initial
ASCs. At each experiment, we randomly generate QoS
attributes and fidelity values of each ASC according to
the following strategy: t is randomly distributed between
10 and 100 ms; c is uniformly distr ibuted [70, 100] $; rel,
m, rep is randomly generated in [0 ,1 ]. The fidelity f or the
first four attributes is in [0,1]. At each experiment point,
we record average values
M
of normalizing matrix M
and average fidelity values
F
of matrix F of selected
ASC with highest score by using Jiang’s algorithm and
our algorithm, respectively, i.e.,
54
,,
1
(), ()
54
ij ij
jj1
M
fid
Mi Fi



(9)
where, i is numbering of ASC with highest QoS per-
formance selected by using Jiang’s algorithm or our al-
gorithm. Let Qei(i) in Formula (10) denote QoS evalua-
tion index for current algorithm, in order to measure final
QoS evaluation of selected ASC. The higher the Qei, the
greater the comprehensive quality performance. In which,
1

and 2

are evaluation weight for QoS attribute and
fidelity factor.
12
()() ()
ei
Qi Mi Fi   (10)
Simulations are divided into three groups for different
weight values of the QoS attributes in Formula (5) con-
sidering different users’ quality preference.
Group 1: , denoting un-
biasedness; 12345
0.2wwwww
Group 2: , showing
preference for t and c;
12 345
0.35, 0.1ww www 
Group 3: 12 showing
preference for other attributes .
345
0.125, 0.25,ww www 
For each group, we run the experiments for different
evaluation weights pair, i.e., 1

from 0.9 down to 0.1,
while 2

from 0.1 up to 0.9. We record average value
of QoS evaluation index for 100 times in each point. The
simulation results are shown in Figure 1, Figure 2, and
Figure 3, respectively.
It can be seen from three figures that our algorithm
gives higher QoS evaluation index performance than
Jiang’s algorithm in most cases for every group. More-
over, we observe from all figures that the performance of
our algorithm is more stable, and the performance of
Jiang’s algorithm decreases as 1

decreases. This is
because that Jiang’s algorithm depends too much upon
QoS values themselves, whereas QoS values and users’
confidence fidelity are better compromised in our algo-
rithm.
1 2 3 4 5 6 7 8 9
0.45
0.5
0.55
0.6
0.65
0.7
0.75
Dif ferent E v aluation W eight s
QoS Evaluation Index
Our propos ed al gori thm
J iang's al gorithm
Figure 1. QoS evaluation index vs weights for Group1
1 2 3 4 5 6 7 8 9
0. 45
0. 5
0. 55
0. 6
0. 65
0. 7
0. 75
Different Eval uation Weights
Q oS E v al uation Index
Our propos ed al gorithm
J iang's al gori t hm
Figure 2. QoS evaluation index vs weights for Group2
1 2 3 4 5 6 7 8 9
0. 45
0.5
0. 55
0.6
0. 65
0.7
0. 75
Different E valuation W eight s
QoS Ev a luation Index
Our propos ed algorithm
J i ang's a lg ori t h m
Figure 3. QoS evaluation index vs weights for Group3

Autonomic Software Component QoS Matchmaking Algorithm Based on Fidelity
Factor in Agent-Based Autonomic Computing System
109
To evaluate the effect of our proposed algorithm fur-
ther, we conduct the following simulations and compare
the QoS evaluation index of our algorithm with Jiang’s
algorithm, Traditional algorithm and Random algorithm.
In traditional algorithm, the first ASC is selected as
match result, i.e., pasc1. Random algorithm randomly
selects ASC. The simulations are also divided into three
groups for different weight values of the QoS attributes
as above. In each group, we assume there are 50 similar
tasks (or ASC requests) and 10 ASC providers. At each
experiment point, we record final QoS evaluation of
selected ASC by using four algorithms, r especti vely.
The simulation results for three groups are shown in
Figure 4, Figure 5, and Figure 6, respectively.
It can be seen from above three figures that our algo-
rithm gives best QoS performance among compared four
algorithms. In almost all experiment request point, the
QoS evaluation index of selected optimal ASC by using
our algorithm is higher than other three algorithms. The
simulation results demonstrate better compromised se-
lection between quality and users’ evaluation in our al-
gorithm.
012345678910 11 1213 14 1516 17 1819 20 21 22 23 2425 26 27 2829 30 31 32 33 3435 36 37 38 39 4041 42 4344 45 4647 48 4950
0
0.2
0.4
0.6
0.8
1
# of experiments
QoS Evaluation Index
Our proposed al gori thm
Jiang algori t hm
Traditi onal al gori t hm
Random al gori thm
Figure 4. A comparison on QoS evaluation index of selected optimal ASC for Group1
0 1 2 34 5 67 8 910 1112 13 1415 16 1718 19 202122 2324 25 262728 293031 32 3334 35 3637 38 3940 4142 43 44 4546 474849 50
0.2
0.4
0.6
0.8
1
# of experiments
QoS E valuat ion I ndex
Our proposed algorithm
Jiang algorithm
Tradi tional al gori thm
Random algori t hm
Figure 5. A comparison on QoS evaluation index of selected optimal ASC for Group2
0123456789101112 131415 16 17 18 19 2021 22 2324 25 262728 293031 32333435 36 3738 39 40 41 4243 44 45 46 47 4849 50
0
0.2
0.4
0.6
0.8
1
# of experiments
QoS E valuation Index
Our proposed algori thm
Jiang algo rithm
Traditional algorit hm
Random al gorit hm
Figure 6. A comparison on QoS evaluation index of selected optimal ASC for Group3

Autonomic Software Component QoS Matchmaking Algorithm Based on Fidelity 110 Factor in Agent-Based Autonomic Computing System
5. Conclusions
In this paper, we discuss the autonomic software com-
ponent QoS matchmaking problem for autonomic ele-
ment, which has been taken as one of the most important
issue in field of autonomic computing, especially in
agent-based autonomic computing system. A QoS model
for ASC is built, and on basis of which, an ASC QoS
matchmaking algorithm based on fidelity factor is pro-
posed. Main work in this paper has following character-
istics:
1) Our proposed QoS model for ASC is simple and
effective, and it does not limit type, amount, and value of
QoS attributes, which shows better scalability and flexi-
bility.
2) During ASC matchmaking process, we introduce
evaluation mechanism of confidence of individual QoS
attributes, i.e., fidelity factor for each attribute, which
can overcome drawbacks such as subjectiveness and
unfairness, and improve the self-configuration capability
for autonomic element.
3) Simulation experiments demonstrate the effective
and correction of our algorithm for matchmaking, and
perform best performance in terms of QoS than other
existing algorithms.
4) Simulations show also that our algorithm has better
compromise between attribute quality and users’ evalua-
tion when selecting ASC. Our proposed algorithm is
suitable for many situations on Internet, such as
e-business, online shopping, in which, people have al-
ways compromised selection need software function
between quality and users’ evaluation.
6. Acknowledgements
This work is supported by the National Natural Science
Foundation of China under Grant No. 90718021.
REFERENCES
[1] J. Kephart and D. Chess, “The vision of autonomic com-
puting,” IEEE Computer Society, No. 1, pp. 41-59, 2003.
[2] N. R. Jennings, “On agent-based software engineering,”
Artificial Intelligence, Vol. 177, No. 2, pp. 277-296, 2000.
[3] G. J. Wickler, “Using expressive and flexible action rep-
resentations to reason about capabilities for intelligent
agent cooperation,” PhD Thesis, University of Edinburgh,
Edinburgh, UK, 1999.
[4] K. Sycara, S. Widoff, M. Klusch, et al., “LARKS: Dy-
namic matchmaking among heterogenous software agents
in cyberspace,” Autonomous Agents and MultiAgent
Systems, Vol. 5, No. 2, pp. 173-203, 2002.
[5] K. Arisha, S. Kraus, F. Ozcan, et al., “IMPACT: The
interactive Maryland platform for agents collaborating
together,” IEEE Intelligent Systems, Vol. 14, No. 2, pp.
64-72, 1999.
[6] Z. L. Zhang and C. Q, Zhang, “An improvement to match-
making algorithms for middle agents,” in AAMAS’02:
Proceedings of the First International Joint Conference on
Autonomous Agents and Multiagent Systems, Bologna,
Italy, pp. 1340-1347, 2002.
[7] Y. C. Jiang and Z. Z. Shi, “Quality of service driven
agent service matchmaking,” Mini-Micro Systems, Vol.
26, No. 4, pp. 687-692, 2005.
[8] R. J. Wang, Y. H. Ru, and X. X. Zhu, “Study on reclaim-
ing agent selection models and approaches,” in IEEE
SOLI’08: Proceedings of IEEE Service Operations and
Logistics, and Informatics, Beijing, China, Vol. 1, pp.
1262-1267, 2008.
[9] Q. Ma, H. Wang, Y. Li, et al., “A semantic QoS-aware
discovery framework for web services,” in IEEE
ICWS’08: Proceedings of IEEE International Conference
on Web Services, Beijing, China, pp. 129-136, 2008.
[10] R. Hu, J. X. Liu, Z. H. Liao, et al., “A web service
matchmaking algorithm based on an extended QoS
model,” in IEEE ICNSC’08: Proceedings of IEEE Inter-
national Conference on Networking, Sensing and Control,
Sanya, China, pp. 1565-1570, 2008.
[11] L. L. Wei, C. L. Chi, M. C. Kuo, et al., “Consumer-centric
QoS-aware selection of web services,” Journal of Com-
puter and System Sciences archive, Vol. 74, No. 2, pp.
211-231, 2008.
[12] E. Giallonardo and E. Zimeo, “More semantics in QoS
matching,” in IEEE SOCA’07: In Proceedings of Interna-
tional Conference on Service-Oriented Computing and
Applications, Newport Beach, USA, pp. 163-171, 2007.
[13] D. K. Guo, Y. Ren, H. H. Chen, et al., “A web services
selection and ranking model with QoS constraints,” Jour-
nal of Shanghai JiaoTong University, Vol. 41, No. 6, pp.
870-875, 2007.
[14] K. J. Liang, Q. Liang, and Y. Yang, “The strategy and
method of QoS parameter processing for grid service
matchmaking,” in ICICIC’06: In Proceedings of First In-
ternational Conference on Innovative Computing, Infor-
mation and Control, Vol. 1, pp. 381-384, 2006.