=Paper=
{{Paper
|id=Vol-1382/paper3
|storemode=property
|title=Recent Possibilities of Intelligent Agents in Distributed Systems
|pdfUrl=https://ceur-ws.org/Vol-1382/paper3.pdf
|volume=Vol-1382
|dblpUrl=https://dblp.org/rec/conf/woa/Wozniak15
}}
==Recent Possibilities of Intelligent Agents in Distributed Systems==
https://ceur-ws.org/Vol-1382/paper3.pdf
Proc. of the 16th Workshop “From Object to Agents” (WOA15) June 17-19, Naples, Italy
Recent Possibilities of Intelligent Agents
in Distributed Systems
Marcin Woźniak∗
∗ Institute of Mathematics, Silesian University of Technology,
Kaszubska 23, 44-100 Gliwice, Poland
Email: Marcin.Wozniak@polsl.pl
Abstract—The article is to discuss recent advances in various Tailored system composition and application of managing
aspects of intelligent agents that perform control functions in agents can improve the overall efficiency [5], [6]. According
workflow management and data processing. Cloud-Computing to research conducted for the European Commission, Cloud-
brings various possibilities of novel approach to data management Computing can decrease amount of money spend on IT by 20%
and efficient computer systems, however the process of distribu- and drastically lower energy consumption. Cloud-Computing
tion must be managed not only to increase efficiency but also
lower energy consumption. This is a task for intelligent agents,
is usually available for users in one of three ways:
that can play crucial role in modern computer science. In this • IaaS (Infrastructure as a Service): special infrastruc-
article recent possibilities for these type of computer systems are
discussed.
ture with dedicated components.
• PaaS (Platform as a Service): special platform of
virtual machines and operation systems.
I. I NTRODUCTION
• SaaS (Software as a Service): users get an access to
Development in computational technology sets demands in applications without integrating into the system, which
front of software engineers. The systems and solutions applied is the most popular model.
in various branches of industry must become more and more
intelligent. These solutions are based on intelligent agents that However mainly some compositions of the mentioned above
have control functions over systems they manage. are more practical. In this article an idea for intelligent
agents managing wrokflows and knowledge distribution will
In modern economy we want computers to manage ef-
be discussed.
ficiently enough not only to increase the income into the
budgets but also to help in environment protection, what means
lower energy consumption and more flexible positioning of A. Related Works
the controlled systems. All these aspects demand intelligent Intelligent Agents assist in processes where computer sys-
technology, that will be able to optimize controlled objects. tems need proper managing. One can point various aspects
Therefore current research concern two important fields of of these situations, however the article is to discuss workflow
computer science: Computational Intelligence and Software management in various systems i.e. with knowledge retrieval.
Engineering. This two, if combined, can give Agents: the
intelligent software and systems. Examinations of distributed systems for workflow man-
agement can be devoted to proper positioning for request
Agents are very important for Cloud-Computing. This management [7], [8], [9], [10]. Other aspects like intelligent
service is becoming more popular in the recent years, because software architectures [11] and efficient management [12] can
of it construction that enables users to use software installed also improve the overall performance. Since new technology
in machines in remote locations. One server can efficiently is still introduced all the time, new possibilities that were not
service many users. The idea for this type of servicing is based available before can be of concern now. Intelligent Agents help
on economic calculations. Mainly it is applied in large corpo- in sorting of big data [13], [14]. Application of intelligent tech-
rations that have many branches placed in remote locations. nologies helps to increase speed of reasoning over incomplete
Second type of Cloud-Computing is available for everybody data and data mining [15], [16]. Image processing also can be
in popular services like document sharing, social networks, improved by advanced software [17], [18], [19], [20].
multimedia streaming and more. First type has mainly eco-
nomic reasons. For a large corporation it is inefficient to have II. I NTELLIGENT W ORKFLOW M ANAGEMENT
similar software departments in every branch. Commonly the
software is used by some workers in different shifts. Moreover Request processing in Cloud-Computing systems (Fig. 1)
if the corporation is international, time zones play an important is commonly managed by two (or more) separate agents. Each
role. While users are working with the system in one branch, client machine sends a request to the system. These requests
somewhere else on the other continent the other team is are ordered by first managing agent and then proceed to next
having a break. Therefore proper configuration of a Cloud- managing agent. This agent composition has a role of the
Computing increase efficient usage of the software, decrease office, similar to offices in authority or local agencies, where
energy consumption what helps to protect environment and each income is indexed and then send to proper department.
improves budget [1], [2], [3], [4]. Similarly in the agent based system, first managing server is
19
Proc. of the 16th Workshop “From Object to Agents” (WOA15) June 17-19, Naples, Italy
Fig. 1. A common model of requests processing in Cloud-Computing system.
Fig. 2. A common model of knowledge processing for Cloud-Computing system.
indexing incoming requests and then sends them to proper (Fig. 2). This processing depends on the requests. Mainly in
agents responsible for different tasks. These agents proceed each type of Cloud-Computing system there is a database (one
indexed incomes using resources or processing units connected o more), where all the information is stored. If a user requests
to the cloud. These type of Cloud-Computing services is used information, a signal is addressed to one of managing agents.
in various verification systems, where requests go through After confirming the rights to obtain requested data, managing
managing agent to the processing server, which is using con- unit starts to proceed. If the agent is asked to provide simple
nection to database to assists verification. Mainly the system is data query the operation is not complicated, from the data
also equipped in backup server, which is to store data backup. base is selected a portion of the data and then it is returned
It ensures that any operation in the system is easily erased and to the user. However if a user requested detailed information
the knowledge can be restored if any demand can happen. In or analysis of the data, very often a new process begins. In
Cloud-Computing systems knowledge is stored in a server that this case managing agent is sending portion of the data for
via connection with other agents is able to process information processing to a dedicated agent which is specialized in this
20
Proc. of the 16th Workshop “From Object to Agents” (WOA15) June 17-19, Naples, Italy
type of operation. In this time user request is indexed and waits are used means of busy period and vacation (idle) time. The
for turn, meanwhile the agent can service another request from explicit formula with detailed information and description for
the queue. This request has similar service. Simple data query conditional joint characteristic functions of τ1 , δ1 and h(τ1 ) is
is processed ad-hoc, other is sent with data to processing agent. presented in [8] and [36]. General equation to calculate these
When processing agent has a result of analysis, it is sent to values is:
managing agent. Here according to indexes in the queue it is
returned to user. All these types of processing can be modeled Bn (s, %, z) = E{e−sτ1 −%δ1 z h(τ1 ) | X(0) = n}, 2 ≤ n ≤ N,
to optimize service costs. (4)
where s ≥ 0, % ≥ 0 and |z| ≤ 1, n ≥ 1. Details on this
equation are discussed in [8], [13] and [36] where using it we
A. Service model
can define components of (3) to model total cost of work:
Workflow managing is important for Quality of Ser-
vice (QoS). Service description for cost optimization defines En e−sτ1 = E{e−sτ1 | X(0) = n} = Bn (s, 0, 1), (5)
Tservice , Tincome and Tvacation , which describe average time then for model traffic finally we have:
of service, average income time and average vacation time
(backup, conservation and etc.). See detailed description in [8]. ∂
En τ1 = − Bn (s, 0, 1) , (6)
Classical cost structure is considered in [21]. While in [22] and ∂s s=0
[23] are presented most important aspects of positioning and
similarly we have:
cost optimization. Various queueing models for applied type of
the server are investigated in [24], [25], [26] and [27]. Please ∂
see also [28] and [29] for a review of important results on En δ1 = − Bn (0, %, 1) . (7)
∂% %=0
modeling and positioning.
Research on similar objects [30], [31], [32], [33], [9] with B. Applied Harmony Search Algorithm
it’s analytical model for traffic are applied in [8]. In the
Harmony Search Algorithm (HSA) was presented in 2001
research a finite-buffer H2 /M/1/N -type QS, similar to server
by Zong Woo Geem. It’s main application is to optimize
traffic modeling functions discussed in [34] and [35] was used.
and position various objects. However the first idea was to
Let it be here presented only a brief description, for details
apply the HSA to compose artificial music. The method in the
please see [8]. Incoming requests describes 2-order distribution
beginning was to perform a music by creating tones according
function:
to the rules of applied music type. The algorithm is using
F (t) = p1 1 − e−λ1 t + p2 1 − e−λ2 t , t > 0, (1) natural adaptation of sounds to compose tones and finally
music theme. It is all based on harmony between sounds
where λi > 0 for i = 1, 2 and p1 , p2 ≥ 0. Inter-arrival times that create the music nice in theme for the people. Musician
are mixed of two exponential distributions with parameters selects the sounds from the scale that best work together. HSA
λ1 and λ2 , which are being “chosen” with probabilities p1 works analogously in the optimization. Each object variable
and p2 . In the system, there are (N − 1) places in queue has a specified range. When choosing the best of the possible
and one for packet in the service. System starts working at numbers to have harmony in the process of optimization we
t = 0 with at least one packet present. After busy period the get the solution.
server begins vacation which is modeled with 2-order hyper
exponential distribution function: HSA can be applied to optimize various objects. If we take
parameters values as the sounds that we must adapt to compose
V (t) = q1 1 − e−α1 t + q2 1 − e−α2 t , t > 0. (2) a music theme we can have efficient optimization method.
HSA will search the space D = [a1 , b1 ] × · · · × [an , bn ]; f :
Interpretation of parameters αi , i = 1, 2 and q1 , q2 is similar D → Rn for optimal values ai ≤ xi ≤ bi ; i = 1, . . . , n that
to that for λi , i = 1, 2 and p1 and p2 . If at the end of vacation optimize the object according to fitness condition f (xi ) →
there is no packet present in the system, the server is on optimum. In it’s simplicity HSA does not demand any special
standby and waits for first arrival to start service process. If restriction for f (·) function. It only must be possible to
there is at least one packet waiting for service in the buffer calculate it’s value at any point of D. In the HSA we define:
at the end of vacation, the service process starts immediately
and new busy period begins. • HM (Harmony Memory) - harmony memory that
Functions F (·) and V (·) help to simulate inter-arrival times stores the
best harmonies used for music composition:
and vacation defined in (1) and (2). In the research optimal x1 = (x11 , . . . , xn1 )|f (x1 )
values for parameters λi , pi , µ and αi are calculated to HM = ...
optimize amount of resources to perform all operations. This xHMS = (x1HM S , . . . , xnHM S )|f (xHMS )
is modeled in rn (c1 ) defined as: As each harmony we understand a set of values
representing positioned object. If the new vector of
Qn (c1 ) r(τ1 )En τ1 + r(δ1 )En δ1 harmony (state of the object) is better than any other
rn (c1 ) = = , (3)
En (c1 ) En τ1 + En δ1 among the previous HM vectors, this new vector
replaces the worst one in the HM. This procedure is
where the symbols are: r(τ1 )-fixed unit operation costs during repeated until the stopping criterion is met,
busy period τ1 , r(δ1 )-fixed unit operation costs during idle
time δ1 , En τ1 -means of busy period τ1 and En δ1 -idle time δ1 • HMS (Harmony Memory Size) - number of harmonies
on condition that system starts with n packets present. In (3) stored in the memory,
21
Proc. of the 16th Workshop “From Object to Agents” (WOA15) June 17-19, Naples, Italy
TABLE I. O PTIMAL PARAMETERS µ, λi , αi , pi , qi FOR i = 1, 2 AND
• HMCR (Harmony Memory Considering Rate) - the LOWEST VALUE OF (3).
coefficient of vector choice for memory in the range
(0, 1). It helps to decide whether, the new component λ1 λ2 α1 α2 p1 p2 q1 q2
will be selected from the memory of harmony HM, 3.1 2.1 1.5 0.3 1.82 1.4 5.8 3.7
µ 0.7 rn (c1 ) 0.32
or will it be the new value of the accepted range of Tservice Tincome Tvacation
variation variables, [sec] 1.42 1.25 16.2
• PAR (Pitch Adjusting Rate) - the tone adjustment
factor in the range (0, 1).
lowest cost of system operation workflow.
Scenario 2.
Algorithm 1 HSA applied to position workflow traffic
Request service is set to Tservice = 2[sec]. Research results
1: Define coefficients: HMS, HMCR, PAR and are shown in Table II.
generations–number of harmony search,
2: Dedicated criterion function: lowest cost of operation (3), TABLE II. O PTIMAL PARAMETERS µ, λi , αi , pi , qi FOR i = 1, 2 AND
LOWEST VALUE OF (3).
3: Create at random initial set HM,
4: t:=0, λ1 λ2 α1 α2 p1 p2 q1 q2
5: while t ≤ generations do 2.4 3.7 0.6 0.5 64.56 0.91 2.6 13.20
6: with probability equal HMCR among all existing har- µ 0.45 rn (c1 ) 0.43
monies in HM xji ∈ xJI , where I = 1, . . . , i, . . . , HM S Tservice Tincome Tvacation
[sec] 2.22 27.14 30.73
and J = 1, . . . , j, . . . , n compose new harmony vector
xnew ,
7: with a probability equal to the value of PAR change
xji = xji + α, where α = b · u and b ∈ [0.01, 0.001] and Scenario 3.
u ∈ [−1, 1], Positioning of heavy traffic i.e. when Tservice = 0.5[sec].
8: with probability equal to 1 - HMCR take randomly new Research results with system positioning are shown in Table
harmony vector x0new variables ai ≤ xi ≤ bi , III.
9: while i ≤ HM S do
TABLE III. O PTIMAL PARAMETERS µ, λi , αi , pi , qi FOR i = 1, 2 AND
10: if f (xnew ) is better then f(x∗ ) then LOWEST VALUE OF (3).
11: change xnew with worst harmony vector x∗ ,
12: end if λ1 λ2 α1 α2 p1 p2 q1 q2
41.3 25.2 107.3 1.8 2.1 1.4 52.21 15.7
13: if f (x0new ) is better then f(x∗ ) then µ 2.34 rn (c1 ) 0.27
14: change x0new with worst harmony vector x∗ , Tservice Tincome Tvacation
15: end if [sec] 0.42 0.1 9.2
16: end while
17: Sort harmonies in HM memory,
18: Next generation: t + +, Scenario 4.
19: end while If we position peculiar incoming traffic Tincome = 2[sec].
20: Best harmony vector in HM memory is potential optimum. Research results are shown in Table IV.
TABLE IV. O PTIMAL PARAMETERS µ, λi , αi , pi , qi FOR i = 1, 2 AND
LOWEST VALUE OF (3).
III. R ESEARCH R ESULTS
λ1 λ2 α1 α2 p1 p2 q1 q2
Application of HSA to traffic modeling will help to position 3.2 4.8 1.11 1.72 5.73 2.92 14.6 6.71
µ 0.43 rn (c1 ) 0.34
operation time and optimize service cost rn (c1 ) for system
Tservice Tincome Tvacation
under-load, critical load or overload. HSA simulations were [sec] 2.32 2.39 17.05
performed for r(τ1 ) = 0.5 and r(δ1 ) = 0.5. It means that
modeled workflow management is simulated for 0.5 energy
unit consumption each vacation and work period. These val-
ues may be changed in (3). Presented modeling results are IV. F INAL R EMARKS
averaged of 100 samplings for HMS = 50 harmonies in 80
generations with HMCR and PAR taken randomly at each In Cloud-Computing or distributed systems where the
generation, where times in the system have equations: amount of data to be transfered over the network is large
optimal managing can significantly influence workflow and
• Average service time: Tservice = µ1 , lower traffic. Presented approach to simulation and positioning
can be a great advantage for distributed systems. Proposed
• Average time between packages income into the sys- positioning and modeling will accelerate operations and it is
tem: Tincome = λp11 + λp22 , not burdened by typical workflow simulation restrictions.
• Average vacation time: Tvacation = αq11 + αq22 , In the article a novel approach to workflow simulation
• Examined system size: N = buffer size +1. and modeling is presented. Proposed novel method is easy
to implement with possibility to improve. Moreover it can
Scenario 1. be implemented in Cloud-Computing where data packages
In Table I are optimum values for all parameters that affect influence workflow stability and performance.
22
Proc. of the 16th Workshop “From Object to Agents” (WOA15) June 17-19, Naples, Italy
R EFERENCES Information Science - ICIST’2014, vol. 465, pp. 312–323, 2014, DOI:
10.1007/978-3-319-11958-8 25.
[1] F. Bonanno, G. Capizzi, G. Lo Sciuto, C. Napoli, G. Pappalardo, and
[18] M. Woźniak and D. Połap, “Basic concept of cuckoo search algorithm
E. Tramontana, “A novel cloud-distributed toolbox for optimal energy
for 2D images processing with some research results : An idea to apply
dispatch management from renewables in igss by using wrnn predictors
cuckoo search algorithm in 2d images key-points search,” in SIGMAP
and gpu parallel solutions,” in Power Electronics, Electrical Drives,
2014 - Proceedings of the 11th International Conference on Signal
Automation and Motion (SPEEDAM), 2014 International Symposium
Processing and Multimedia Applications, Part of ICETE 2014 - 11th
on. IEEE, 2014, pp. 1077–1084.
International Joint Conference on e-Business and Telecommunications.
[2] F. Bonanno, G. Capizzi, and C. Napoli, “Some remarks on the appli- 28-30 August, Vienna, Austria: SciTePress, 2014, pp. 157–164, DOI:
cation of rnn and prnn for the charge-discharge simulation of advanced 10.5220/0005015801570164.
lithium-ions battery energy storage,” in Power Electronics, Electrical
[19] C. Napoli, G. Pappalardo, E. Tramontana, Z. Marszałek, D. Połap,
Drives, Automation and Motion (SPEEDAM), 2012 International Sym-
and M. Woźniak, “Simplified firefly algorithm for 2D image key-
posium on. IEEE, 2012, pp. 941–945.
points search,” in IEEE SSCI 2014 - 2014 IEEE Symposium Series on
[3] G. Capizzi, F. Bonanno, and C. Napoli, “Recurrent neural network- Computational Intelligence - CIHLI 2014: 2014 IEEE Symposium on
based control strategy for battery energy storage in generation systems Computational Intelligence for Human-Like Intelligence, Proceedings.
with intermittent renewable energy sources,” in International conference 9-12 December, Orlando, Florida, USA: IEEE, 2014, pp. 118–125, DOI:
on clean electrical power (ICCEP). IEEE, 2011, pp. 336–340. 10.1109/CIHLI.2014.7013395.
[4] G. Capizzi, F. Bonanno, and C. Napoli, “A new approach for lead-acid [20] M. Woźniak, D. Połap, M. Gabryel, R. K. Nowicki, C. Napoli, and
batteries modeling by local cosine,” in Power Electronics Electrical E. Tramontana, “Can we preprocess 2d images using artificial bee
Drives Automation and Motion (SPEEDAM), 2010 International Sym- colony?” Lecture Notes in Artificial Intelligence - ICAISC’2015, vol.
posium on. IEEE, 2010, pp. 1074–1079. 9119, pp. 660–671, 2015, DOI: 10.1007/978-3-319-19324-3 59.
[5] C. Napoli, G. Pappalardo, E. Tramontana, and G. Zappalà, “A cloud- [21] J. Teghem, “Control of the service process in a queueing system,”
distributed gpu architecture for pattern identification in segmented Europ. Jour. of Operations Research, vol. 1, no. 23, pp. 141–158, 1986.
detectors big-data surveys,” The Computer Journal, p. bxu147, 2014. [22] O. Kella, “Optimal control of the vacation scheme in an M/G/1
[6] C. Napoli, G. Pappalardo, and E. Tramontana, “Improving files avail- queue,” Operations Research Journal, vol. 4, no. 38, pp. 724–728, 1990.
ability for bittorrent using a diffusion model,” in 23rd IEEE Interna- [23] R. Lillo, “Optimal operating policy for an M/G/1 exhaustive server-
tional WETICE Conference. IEEE, 2014, pp. 191–196. vacation model,” Methodology and Computing in Applied Probability,
[7] M. Gabryel, M. Woźniak, and R. Damaševic̆ius, “An application of vol. 2, no. 2, pp. 153–167, 2000.
differential evolution to positioning queueing systems,” Lecture Notes [24] U. C. Gupta, A. D. Banik, and S. Pathak, “Complete analysis of
in Artificial Intelligence - ICAISC’2015, vol. 9120, pp. 379–390, 2015, M AP/G/1/N queue with single (multiple) vacation(s) under limited
DOI: 10.1007/978-3-319-19369-4 34. service discipline,” Journal of Applied Mathematics and Stochastic
[8] M. Woźniak, W. M. Kempa, M. Gabryel, and R. K. Nowicki, “A Analysis, no. 3, pp. 353–373, 2005.
finite-buffer queue with single vacation policy - analytical study with [25] U. C. Gupta and K. Sikdar, “Computing queue length distributions in
evolutionary positioning,” International Journal of Applied Mathematics M AP/G/1/N queue under single and multiple vacation,” Applied
and Computer Science, vol. 24, no. 4, pp. 887–900, 2014, DOI: Mathematics and Computation, vol. 2, no. 174, pp. 1498–1525, 2006.
10.2478/amcs-2014-0065.
[26] Z. Niu and Y. Takahashi, “A finite-capacity queue with exhaus-
[9] M. Woźniak, “On positioning traffic in nosql database systems by the tive vacation/close-down/setup times and markovian arrival processes,”
use of particle swarm algorithm,” in Proceedings of XV Workshop Queueing Systems, vol. 1, no. 31, pp. 1–23, 1999.
DAGLI OGGETTI AGLI AGENTI - WOA’2014, vol. 1260. 25-
26 September, Catania, Italy: CEUR Workshop Proceedings (CEUR- [27] Z. Niu, T. Shu, and Y. Takahashi, “A vacation queue with setup and
WS.org), RWTH Aachen University, 2014. close-down times and batch markovian arrival processes,” Performance
Evaluation Journal, vol. 3, no. 54, pp. 225–248, 2003.
[10] M. Woźniak, “On applying cuckoo search algorithm to positioning
GI/M/1/N finite-buffer queue with a single vacation policy,” in [28] H. Takagi, Queueing Analysis, vol. 1: Vacation and Priority Systems,
Proceedings of the 12th Mexican International Conference on Artificial vol. 2. Finite Systems. Amsterdam: North-Holland, 1993.
Intelligence - MICAI’2013. 24-30 November, Mexico City, Mexico: [29] N. Tian and Z. G. Zhang, Vacation queueing models. Theory and
IEEE, 2013, pp. 59–64, DOI: 10.1109/MICAI.2013.12. applications. Berlin, Heidelberg: Springer - Verlag, 2006.
[11] G. Borowik, M. Woźniak, A. Fornaia, R. Giunta, C. Napoli, G. Pap- [30] W. M. Kempa, “GI/G/1/∞ batch arrival queuing system with a single
palardo, and E. Tramontana, “A software architecture assisting workflow exponential vacation,” Mathematical Methods of Operations Research,
executions on cloud resources,” International Journal of Electron- vol. 1, no. 69, pp. 81–97, 2009.
ics and Telecommunications, vol. 61, no. 1, pp. 17–23, 2015, DOI: [31] W. M. Kempa, “Characteristics of vacation cycle in the batch arrival
10.1515/eletel-2015-0002. queuing system with single vacations and exhaustive service,” Int. Jour.
[12] G. Ćwikła, A. Sekala, and M. Woźniak, “The expert system supporting of Applied Mathematics, vol. 4, no. 23, pp. 747–758, 2010.
design of the manufacturing information acquisition system (MIAS) for [32] W. M. Kempa, “Some new results for departure process in the
production management,” Advanced Materials Research, vol. 1036, pp. M X /G/1 queuing system with a single vacation and exhaustive
852–857, 2014, DOI: 10.4028/www.scientific.net/AMR.1036.852. service,” Stochastic Analysis and Applications, vol. 1, no. 28, pp. 26–43,
[13] M. Woźniak, Z. Marszałek, M. Gabryel, and R. K. Nowicki, “Modified 2009.
merge sort algorithm for large scale data sets,” Lecture Notes in [33] W. M. Kempa, “On departure process in the batch arrival queue with
Artificial Intelligence - ICAISC’2013, vol. 7895, pp. 612–622, 2013, single vacation and setup time,” Annales UMCS Informatica, vol. 1,
DOI: 10.1007/978-3-642-38610-7 56. no. 10, pp. 93–102, 2010.
[14] M. Woźniak and Z. Marszałek, Extended Algorithms for Sorting Large [34] D. Hongwei, Z. Dongfeng, and Z. Yifan, “Performance analysis of
Data Sets. Poland: Silesian University of Technology Press, 2014. wireless sensor networks of serial transmission mode with vacation on
[15] R. K. Nowicki, B. Nowak, and M. Woźniak, “Rough k nearest neigh- fire prevention,” ICCET’10 IEEE CPS, pp. 153–155, 2010.
bours for classification in the case of missing input data,” in Proceedings [35] V. Mancuso and S. Alouf, “Analysis of power saving with continuous
of the 9th International Conference on Knowledge, Information and connectivity,” Comp. Networks, vol. 56, no. 10, pp. 2481–2493, 2012.
Creativity Support Systems, G. A. Papadopoulos, Ed. 6-8 November, [36] M. Woźniak, W. M. Kempa, M. Gabryel, R. K. Nowicki, and Z. Shao,
Limassol, Cyprus: University of Cyprus Press, 2014, pp. 196–207. “On applying evolutionary computation methods to optimization of
[16] B. Nowak, R. K. Nowicki, M. Woźniak, and C. Napoli, “Multi-class vacation cycle costs in finite-buffer queue,” Lecture Notes in Artifi-
nearest neighbor classifier for incomplete data handling,” Lecture Notes cial Intelligence - ICAISC’2014, vol. 8467, pp. 480–491, 2014, DOI:
in Artificial Intelligence - ICAISC’2015, vol. 9119, pp. 465–476, 2015. 10.1007/978-3-319-07173-2 41.
[17] M. Woźniak and Z. Marszałek, “An idea to apply firefly algorithm
in 2D images key-points search,” Communications in Computer and
23