Docstoc

IMPLEMENTATION OF PARTICLE SWARM OPTIMIZATION _PSO_ ALGORITHM ON POTATO EXP

Document Sample
IMPLEMENTATION OF PARTICLE SWARM OPTIMIZATION _PSO_ ALGORITHM ON POTATO EXP Powered By Docstoc
					 INTERNATIONAL JOURNAL OF COMPUTER ENGINEERING &
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 4, July-August (2013), © IAEME
                             TECHNOLOGY (IJCET)

ISSN 0976 – 6367(Print)
ISSN 0976 – 6375(Online)                                                       IJCET
Volume 4, Issue 4, July-August (2013), pp. 82-90
© IAEME: www.iaeme.com/ijcet.asp
Journal Impact Factor (2013): 6.1302 (Calculated by GISI)                   ©IAEME
www.jifactor.com




       IMPLEMENTATION OF PARTICLE SWARM OPTIMIZATION (PSO)
              ALGORITHM ON POTATO EXPERT SYSTEM

                      A.Sri Rama Chandra Murty1& M. Surendra Prasad Babu2
 1&2
       (Dept. of Computer Science & Systems Engineering, Andhra University, Vishakhapatnam, A.P)



ABSTRACT

       The aim of this paper is to providing expert advice & information to the potato cultivators by
developing Potato Expert System. A Particle Swarm Optimization is evolutionary computational
technique which can be applied to solve optimization problems. The Potato Expert System is
implementation by using PSO algorithm in three phases: first, developing Potato Knowledge base.
Second, Machine learning algorithm is implemented for data collection. Third, Potato Expert System
shell developed using Rule Based Expert System with Backward Chaining. The Potato Expert
System interface tool is developed that provides experts’ advice to cultivators for disease
management.

Keywords: Particle Swarm Optimization, Potato Expert System, Machine Learning, Genetic
Algorithms, Backward Chaining, Rule Based Expert System.

I.        INTRODUCTION

        Expert system in general simulates both the knowledge and know-how of human experts i.e.,
the expert system solves problems that are normally solved by human experts. All expert systems
include at least three basic elements i.e., a knowledge base, which represents, what is known about a
given subject at present, an interface engine comprises the logistics to apply what is known to what is
yet unknown. Expert systems are most common in specific problem domain, and are traditional
application and/or subfield of artificial intelligence.
        Based on the different representation schemes and interface techniques, the expert systems
are classified as rule based expert systems, frame-based expert systems case based expert system and
model based expert system. Several representations such as list of rules and facts, frames and slots,
semantic networks, OAV- triplets etc., are used in the above expert system. The interference engine
may infer conclusion from the knowledge base and the fact supplied by the user. Several expert
systems are widely used in various fields like medicines, geology, system configuration and
engineering design and structural analysis of chemical compounds.

                                                  82
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 4, July-August (2013), © IAEME

        A rule-based, expert system maintains a separation between its Knowledge base and that part
of the system that executes rules, which was often referred to as the expert system shell. Rule-based
system represents knowledge as a bunch of rules and assertions. It involves a database that stores the
assertions and rules that can perform some action (production system) or deduce consequences
(deduction systems). The implementation of control over a finite set of assertions allow the systems
to dynamically generate new knowledge (forward chaining) and breakdown a complicated problems
in to many smaller ones (backward chaining). Unification (matching variables) allow flexibility of
the rules, with which the systems can deduce more specific facts (forward chaining) and solve more
specific problems (backward chaining) without having a giant set of smaller rules. Deduction
systems contain only IF-THEN rules.
        Machine Learning refers to a system which is capable of autonomous acquisition and
integration of knowledge. The goal of machine learning is to program computers to use example data
or past experience to solve a given problem. Many successful applications of machine learning exist
already, including systems that analyse past sales data to predict customer behaviour, recognize faces
or spoken speech, optimize robot behaviour so that a task can be completed using minimum
resources, and extract knowledge from bio-informatics data.
        Potato (Solanum tuberosum L.) popularly known as ‘The king of vegetables’, has emerged as
fourth most important food crop in India after rice, wheat and maize. Indian vegetable basket is
incomplete without Potato. Because, the dry matter, edible energy and edible protein content of
potato makes it nutritionally superior vegetables as well as staple food not only in our country but
also throughout the world. Hence, potato may prove to be a useful tool to achieve the nutritional
security of the nation.
        Several varieties are grown in different parts of India. China and Rio-De- Janeiro are the two
imported varieties of potato. Other important varieties grown are kufri, Chipsona-1, Kufri Chipsona-
2, Kufri Jyothi, Kufri Luvkar, Kufri Chandramukhi, Kufri Badsah, Kufri Lavkar, Kufri Lalima, Kufri
Sindhuri, Irish Cobbler, Red Pontiac, Viking, Katahdin etc. This information is stored in Knowledge
base and retrieved through information system module. Several programmed interviews were
conducted with agricultural experts & progressive farmers and identified different diseases likes
Black Scurf, Potato Wart, Common Scab, Late Blight, Early Blight, Black leg and Soft Rot etc.,

II.    POTATO EXPERT SYSTEM ARCHITECTURE

        The architecture of the proposed Potato Expert System consists of 3 modules: (1) User
Interface (2) Interface Engine (3) Advices from Potato Expert System. The user interacts with the
system through a specially designed unified interface which assimilates the peculiarities of the
various components. A graphical user interface (GUI) provides a user friendly and comfortable
environment in which he/she works and communicates with Potato Expert System.




                         Fig.1. Proposed Potato Expert System Architecture

                                                 83
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 4, July-August (2013), © IAEME

        The GUI presents interactive forms and command menus to retrieve and update system
parameters and steering variables, to enter user constraints and preferences and to prevent relevant
diagnosis, treatment information back to the user after simulations have run Potato Expert users can
query the system using an inference process that automatically matches facts against patterns and
determines which rules are applicable. When calling for the results of the diagnosis, the system will
explain the inference processes. After an examination of the facts collected from users, the system
will produce conclusions of the diagnosis and treatment methods. If inference process not matches
facts against the machine learning algorithm will be executed and produces the diagnosis and
treatment methods. The Knowledge base contains all the rules for the fish disease diagnosis. Each
rule has two sections – a symptom pattern section and an action section, in the form of ‘IF symptom
pattern E, THEN the disease H’. The advices produced by the expert system displays on the output
screen.

III.   KNOWLEDGE BASE POTATO EXPERT SYSTEM

The rules used for the rule based system are extracted from Table 1.2 are given below.
Rule 1: IF the stage of the crop is seeding and the part of the crop affected is leaves and light
yellowing of the tips of lower leaves THEN the disease is Soft rot.
Rule 2: IF the stage of the crop is seedling and part of the crop affected is leaves and very small
round scattered spots in the youngest levels which increases with plant growth THEN the disease is
Leaf spot.
Rule 3: IF the stage of the crop is seedling and the part of the crop affected is leaves and leaves of
infected plants tend to be narrower and more erect THEN the disease is Rhizome rot.
Rule 4: IF the stage of the crop is seedling and the part of the crop affected is leaves and small
powdery pustules present over both surfaces of the leaves THEN the disease is White grub.
Rule 5: IF the stage of the crop is seedling and the part of the crop affected is leaves and lesions
begin as small regular elongated necrotic spots and grow parallel to the veins THEN the disease is
Gray leaf spot.
Rule 6: IF the stage of the crop is seedling and the part of the crop affected is leaves and lesions with
oval narrow necrotic and parallel to the veins THEN the disease is Phyllosticta leaf spot.
Rule 7: IF the stage of the crop is seedling and the part of the crop affected is root white thin lesions
along leaf surface and green tissue in plants THEN the disease is Thread blight.
Rule 8: IF the stage of the crop is seedling and the part of the crop affected is root and the bushy
appearance due to proliferation of tillers which become chlorotic, reddish and lodging THEN the
disease is Fusarial wilt.
Rule 9: IF the stage of the crop is seedling and the part of the crop affected is root and irregular
section of epidermis and perforated leaves THEN the disease is Dry rot.
Rule 10: IF the stage of the crop is seedling and the part of the crop affected is stem and the affected
area just above the soil line is brown water-soaked soft and collapsed THEN the disease is Bacterial
wilt.
Rule 11: IF the stage of the crop is seedling and the part of the crop affected is stem and affected
internodes become disintegrated and the presence of small pin-head like black sclerotic on the rind of
the stalks THEN the disease is Rhizome scale.
Rule 12: IF the stage of the crop is seedling and the part of the crop affected is stem and wilted
plants remain standing when dry and small dark-brown lesions develop in the lowest internodes
THEN the disease is Mosaic streak.
The following Database Records in Table 1 is used in the implementation of Rule Based Expert
System.


                                                   84
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 4, July-August (2013), © IAEME




                  TABLE.1. Rule Based Expert System Database Records




             TABLE.2. Machine Learning Symptoms Database Table Description

                                          85
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 4, July-August (2013), © IAEME

       In machine learning system, the rules are stored in the form of 1’s and 0’s which are YES/NO
condition of the symptoms. These rules with corresponding diseases and their cure information are
represented below in Table 2, 3 & 4.




                TABLE.3. Machine Learning Decision Database Table Description




                     TABLE.4. Machine Learning Database Table Description

IV.    RULE-BASED EXPERT SYSTEM USING BACKWARD CHAINING

       The first step in developing backward chaining system is to define the problem or learn about
the problem through reports, documents and books. The second step is to define the goals for the
system. Every backward chaining system needs at least one goal to get started. By defining the goals
this may help us to start from the right track and end on the expected track and this may avoid us
                                                86
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 4, July-August (2013), © IAEME

from being misled from the real problem. The third step is designing the goal rules, such as IF
precondition1 AND precondition2 THEN portfilio1. With the goal portfolio1 is attained but undergo
first the two preconditions, the goal rules also have decision table to make and help the decision
making with this rules and testing for the goal rules.
         The fourth step is to expand the knowledge of the system and these expansion techniques are
broadening the system knowledge that teaches the system about additional issues and deepening the
system knowledge teaches about the issues it already known. The fifth step is refining the system in
which there are several additional features that will enhance both its performance and maintenance.
The sixth step is designing the interface that will accommodate the needs of the user so that the user
can choose easily and efficiently. The final step is system evaluation and this is done by making
some questions to the expert and tests the system with sample input and sees if the system is really
running properly. The Backward chaining Mechanism for Rules based Potato Expert System is
shown in Fig.2.




            Fig.2. Backward Channing Mechanism for Rule Based Potato Expert System

V.     DESIGN OF PSO ALGORITHM FOR POTATO EXPERT SYSTEM

        Particle Swarm Optimization (PSO) [1][2][3] is a population based stochastic optimization
technique, inspired by social behaviour of bird flocking or fish schooling. PSO shares many
similarities with evolutionary computation techniques such as Genetic Algorithms (GA) [4].
However, unlike GA, PSO has no evolution operators such as crossover and mutation. In PSO, the
potential solutions, called particles, fly through the problem space by following the current optimum
particles. Each particle keeps track of its coordinates in the problem space which are associated with
the best solution (fitness) it has achieved so far. This value is called pbest [5]. Another best value that
is tracked by the particle swarm optimizer is the best value, obtained so far by any particle in the
neighbours of the particle. This location is called lbest [6].
        When a particle takes all the population as its topological neighbours, the best value is a
global best and is called gbest. The particle swarm optimization concept consists of, at each time
step, changing the velocity of (accelerating) each particle towards its pbest and lbest locations.

                                                    87
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 4, July-August (2013), © IAEME

Acceleration is weighted by a random term, with separate random numbers being generated for
acceleration towards pbest and lbest locations.




                     Fig.3. Proposed PSO Architecture for Potato Expert System

         PSO simulates the behaviours of bird flocking. Suppose the following scenario: a group of
birds are randomly searching food in the area. There is only one piece of food in the area being
searched. All the birds do not know where the food is. But they know how far the food in every
iteration. So what’s the best strategy to find the food? The effective one is to follow the bird which is
nearest to the food. PSO learned from the scenario and used it to solve the optimization problems. In
PSO, each single solution is a “bird” in the search space. We call it “particle”. All of particles have
velocities which direct the flying of the particles.
         The particles fly through the problem space by following the current optimum particles. PSO
is initialized with a group of random particles (solutions) and then searches for optima by updating
generations. Every particle is updated by two best values in all iterations. The first best solution is
the fitness value called pbest. Another best value that is tracked by the particle swarm optimizer and
obtained so far by any particle in the population is the global best value called gbest. When the
particle takes part of the population as its topological neighbours the best value is the local best and
is called lbest.
         After finding the two best values, the particle update its velocity and positions with following
equations
         v[]=v[] +c1 * rand() * (pbest[] – present[])+c2 * rand() * (gbest[] – present[])… (a)
         present[] = present[] + v[] ….                                                         (b)
v[] is the particle velocity, present[] is the current particle (solution). Pbest[] and gbest[] are defined
as stated before. rand() is a random number between (0,1). c1,c2 are learning factors, usually
c1=c2=2.
Procedure of Proposed PSO Algorithm
Step 1: For each particle initialize.
Step 2: For each particle calculate fitness value.
Step 3: If the fitness value is better than the best fitness value (pbest) in history set current value as
the new pbest.
Step 4: Choose the particle with the best fitness value of all the particles as the gbest.
Step 5: For each particle calculate particle velocity ie., equation (a)
Step 6: Update the particle position ie., equation (b).
         While maximum iterations or minimum error criteria is not attained, particles velocities on
each dimension are clamped to a maximum velocity Vmax. If the sum of accelerations causes the
velocity on that dimension to exceed Vmax, a parameter specified by the user, then the velocity on
that dimension is limited to Vmax. However, PSO does not have genetic operators like crossover and
mutation. Particles update themselves with the internal velocity. Compared with GAs the

                                                    88
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 4, July-August (2013), © IAEME

information sharing mechanism in PSO is significantly different. In GAs, chromosomes share
information with each other. So the whole population moves like a one group towards an optimal
area. In PSO, only gbest (or lbest) gives out the information to others. It is one-way information
sharing mechanism. The evolution only looks for the best solution. Compared with GA, all the
particles tend to converge to the best solution quickly even in the local version in most cases.

VI.    CONCLUSIONS AND FUTURE WORK

        The main emphasis of the paper is to have a well-designed interface for giving Potato plant
related advices and suggestions in the area to farmers by providing facilities like online interaction
between expert system and the user without the need of expert all the time. based on the proposed
architecture, a new expert system shell, capable of developing a Rule based Expert System with
Machine Learning capabilities, especially using PSO algorithm, will be designed. The expert systems
developed using this shell are expected to show better performance than the systems developed using
other algorithms like ABC algorithm and ACO algorithm.
        The proposed interface tool can be extended further by creating more features and facilities to
the user and the subject expert. The shell can include all the system designs like static and dynamic
systems and other user friendly features so that the expert can design and make any changes online to
the system according to the future R&D in the crop production through proper administration
privileges.
        The present system is developed for only disease management of potato crop, to provide a
complete advice to potato farmers; it is to be extended to all the aspects in farming such as Soil
management, Fertilizer management, Irrigation management, Marketing & Storage management,
Crop management etc. The system may be improved much by adding new features like language
translation facilities and adding the new updates in the crops and production techniques and
embedding audio, video clips and IVRS systems.

REFERENCES

[1]    Kennedy, J and Eberhart, R.C. Particle Swarm Optimization. Proc. IEEE Int’l conf. on Neural
       Networks Vol. IV, pp. 1942-1948. IEEE Service centre, Piscataway, NJ, 1995.
[2]    Eberhart, R. C. and Kennedy, J. A new Optimizer using Particle Swarm Theory. Proc. of the
       sixth International Symposium on micro machine and human science pp. 39-43. IEEE
       Service Centre, Piscataway, NJ, Nagoya, Japan, 1995.
[3]    Eberhart, R. C. and Shi, Y. Particle Swarm Optimization: developments, applications and
       resources. Proc. congress on evolutionary computation 2001 IEEE service centre, Piscataway,
       NJ, Seoul, Korea., 2001.
[4]    Eberhart, R. C. and Shi, Y. Evoluting artificial Neural Networks. Proc. 1998 Int’l Conference
       on neural networks and begon pp. PL5 – PL13, Beijing, P.R. China, 1998.
[5]    Eberhart, R. C. and Shi, Y. Comparison between genetic algorithms and particle swarm
       optimization. Evolutionary programming vii: Proc. 7th ann. Conf. on evolutionary conf.,
       Spring-Verlag, Berlin, San Diego, CA., 1998.
[6]    Shi, Y. and Eberhart, R. C. Parameter selection in Particle Swarm Optimization. Evolutionary
       Programming VII: Proc. EP 98 pp 591-600.
[7]    Shi. Y. and Ebernard, R. C. A modified Particle Swarm Optimizer. Proc. of the IEEE and
       International Conference on Evolutionary Computation pp. 69-73, IEEE Press, Piscataway,
       NJ, 1998.
[8]    Holland, J. H. (1992). Adaptation in Neural and Artificial Systems. MIT Press, Cambridge
       MA.

                                                  89
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 4, July-August (2013), © IAEME

[9]    Millonas, M. M. (1994). Swarms, Phase transisions, and collective intelligence. In C. G. Langton,
       Ed., Artificial Life III, Addision wesely, Reading, M.A.
[10]   Reeves, W.T. (1983) Partcile systems – a technique for modelling a class of fuzzy objects. ACM
       Transations on Graphics. 2(2):91-108.
[11]   Reynolds, C. W.(1987). Flocks, herds and schools: a distributed behavioural model. Computer
       Graphics. 21(4):25-34.
[12]   Oliver Wirjadi, Thomas M. Beeuel, Wolfgang Feiden , and Yoo-jin kim. Automated Feature
       Selection for the Classification of Meningioma cell Nuclei, Institute of Neuropathology, 2006.
[13]   Huajin, “Classification Algortithm for the Tip fracture prediction based on Recursive Partitioning
       Methods”, Dept. of Radiology, University of California, San Francisco, South China, 2004.
[14]   Uzeyir, Gurbanli, “Application of Analysis Methods in Risk Management”, Institute of
       Information Technologies, Baku,, Azerbaijan.
[15]   Juraiza Ishak, Ansor, Md Tahir, Nooritawati Hussan, Aini Mustafa, Mohd Marzuki, “Weed
       Classification using Decision Tree”, International Symposium, 2, issue 26-28, page 1-5, Aug
       2008.
[16]   R. Arivoli and Dr. I. A. Chidambaram, “Multi-Objective Particle Swarm Optimization Based
       Load-Frequency Control of a Two-Area Power System with Smes Inter Connected using Ac-Dc
       Tie-Lines”, International Journal of Electrical Engineering & Technology (IJEET), Volume 3,
       Issue 1, 2012, pp. 1 - 20, ISSN Print : 0976-6545, ISSN Online: 0976-6553.
[17]   Adesh Chandra, Anurag Singh and Ishan Rastogi, “Understanding Enterprise Risk Management
       and Fair Model with the Help of a Case Study”, International Journal of Computer Engineering &
       Technology (IJCET), Volume 3, Issue 3, 2012, pp. 300 - 311, ISSN Print: 0976 – 6367, ISSN
       Online: 0976 – 6375.

BIOGRAPHIES

                  A.Sri Rama Chandra Murty, (Andhra Pradesh Civil Services – Executive
                  Branch) Special Grade Deputy Collector and presently working as Executive
                  Officer, Srikalahasteeswara Swamy Vari Devasthanam, Srikalahasti (A.P). He
                  obtained B.Com., M.Tech. (Computer Science & Technology – Artificial
                  Intelligence & Robotics), M.Sc. (Information Technology), MCA, PGDCPA,
                  B.L., L.L.M. (Corporate & Security laws), MBA (Finance & Marketing dual
                  specialization), MBA (Human Resources), M.Sc. (Psychology), M.A. (Public
Administration), PGDIRPM, MSPR (Master of Science in Public Relations). His areas of interests
are Expert systems and Artificial Intelligence.



                    M. Surendra Prasad Babu obtained his B. Sc, M.Sc. and M. Phil and
                   Ph.D. degrees from Andhra University in 1976, 1978, 1981and 1986
                   respectively. During his 27 years of experience in teaching and research, he
                   attended about 28 National and International Conferences/ Seminars in India and
                   contributed about 33 papers either in journals or in National and International
                   conferences/ seminars. He has guided 98 student dissertations of B.E., B.Tech.
                   M.Tech. & Ph.Ds. He is now Head of the Department of Computer Science &
                   Systems Engineering of Andhra University College of Engineering, Andhra
University, Visakhapatnam. He received the ISCA Young Scientist Award at the 73rd Indian Science
Congress in 1986 from the hands of late Prime Minister Shri Rajiv Gandhi.



                                                  90

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:1
posted:7/26/2013
language:Unknown
pages:9