Monday, December 14, 2009

Roultte Wheel Selection in Genetic Algorithm

Roulette Wheel was one of most common method among selection methods in GA. It start after generate population of chromosome work has been completed. The problem is how to select individual population in crossovering method. Before selecting each, every individual population need to calculate their fitness and rank it by good solution in higher and lower for less good solution. Then it can represent by a big round cake or circle. The big partition cake for good solution and small partition cake for less good solution. The idea was every good solution will have more probability compare with less good solution. This picture below are taken from others website. Population no. 2 have small shape because of less good solution compare with no. 2 and no.3 where it was more good solution chromosome.
After that throw the dice to get random number, then two selection from population cromosome for crossover process. It also can be use to select which population that can be use for mutation process. Finally we can select the best solution among every population, crossover and mutation process. These idea for fast searching solution compare with conventional method. Thank you for reading.

Saturday, December 5, 2009

Encoding Problem into Genetic Algorithm process


When dealing with Genetic Algorithm optimization method, there is need to encode problem that can be suite in Genetic Algorithm process.
In last article I have mention some about few ways to encode problem. Now I just want to explain more about one of the method that what I be use which is Permutation Encoding.
Permutation encoding can be used in ordering problems, such as traveling salesman problem or task ordering problem. In permutation encoding, every chromosome is a string of numbers, which represents number in a sequence.
In my research I use permutation where it can give order chosen node to be sleep mode, than monitor the delay on network as it fitness. Time slot of node to be sleep mode will be change for each chromosome population if packet delay of network increase. The process will be repeated until one solution have been found.
  • Crossover
Single point crossover - one crossover point is selected, till this point the permutation is copied from the first parent, then the second parent is scanned and if the number is not yet in the offspring it is added
(1 2 3 4 5 6 7 8 9) Chromosome A +
(4 5 3 6 8 9 7 2 1) Chromosome B
=> (1 2 3 4 5 6 8 9 7) After crossover
  • Mutation
Order changing - two numbers are selected and exchanged
(1 2 3 4 5 6 8 9 7) => (1 8 3 4 5 6 2 9 7)

Thursday, December 3, 2009

Process involve in Genetic Algorithm

Hai, today I would like to tell more detail about process that need to be understand when dealing with Genetic Algorithm. Before that, there is need to encoding the problem to be represent by individual cromosome, so that can be process in Genetic Algorithm. There are four main type that can be represent of problem situation. There are:
  1. Binary Encoding
  2. Permutation Encoding
  3. Value Encoding
  4. Tree Encoding
After complete identify which type encoding need to be choosen, then can go through futher prosess. There are four main processes there are:

  1. Selection
  2. Crossover
  3. Mutation
  4. Regeneration

1. Selection

Under this process also have 3 methods where:

  1. Elitism
  2. Roulette Wheel
  3. Tournament Selection

2. Crossover

Under this process also have many methods but 2 of them are:

  1. Exponential Scaling
  2. Linear Normalize

3. Mutation

Under this process just similar between all researchers that using GA

4. Regeneration

Under this process it will select the best population for next iteration, so if there are new child cromosome that better with their parent cromosome from previous generation then it will replace or overwrite that parent cromosome.

Thursday, November 19, 2009

Formula for GA iteration

Today I understand how Genetic Algorithm work where it can give solution. It start with a problem where we need to identify how we can come out with solution. I have some guideline where to start with:
  1. Encoding of the problem in binary string.
  2. Random generating of a population. This one includes a genetic pool representing a group of possible solution.
  3. Reckoning of a fitness value for each subject. It will directly depend on the distance of optimum.
  4. Selection of the subjects that will mate according to their share in the population global fitness.
  5. Genomes crossover and mutations.
  6. And then start again from point 3.

For now I need to try some example that related with GA. My focus now is understanding the GA operation itself and to know how GA can help me using optimization method to find my solution. Example each node that I need to force to sleep mode will represent by a bit. Bit 1 represent a node in active mode and bit 0 will represent node that in sleeping mode. With this all bit representative can build one chromosome. With this chromosome it can start with GA operation that will come with new offspring. This new offspring can be test with a formula or simulation that can give result where it was good chromosome or bad chromosome.

Wednesday, November 4, 2009

First day in NTU

.

It was great to foot step myself in NTU. First thing need to do was preparing the slide presentation for brieft introduction my research to NTU supervisor and suddently he want to change the title for attachment project, so that it can suite with my research. It is good thing because I dont need to waste my time to study other algorithm that not related with my research. Straight away do my research here with under attachment supervisor monitoring. Here some picture that shooting during walking to ministry of man power, Singapore.
.
.

Tuesday, October 20, 2009

ICSECS'09 Conference Presentation

.AlhamduliLah today I was attending ICSECS09 conference that held at Swiss Garden & Spa, Kuantan, Pahang and it was my 4'th conferences that I have attended. I quite happy despite Alex, Ira & also master sifu Sir Suhaimi beside others researcher from other places in others background to participate in here. Everything work so fine and with Mr Suhaimi always entertain joke & tips that make me stay calm before my presentation. Very thank you to UMP which is the organizer for great hospitality.

.

Sunday, September 6, 2009

Confirmation of acceptance from NTU Univ.

Last week I received a letter from Chair of Research of NTU, wrote like this:

Dear Mohd Taufiq

Attached please find the scan copy of the approved letter.
The Original Copy has been sent to you via normal airmail.
Please keep a close look out then.

regards

Ms Len Ah Chan
for A/P Lee Bu Sung, Francis
Associate Chair (Research)


This is my great opportunity to study at oversea, I mean even do Singapore is just very near to Malaysia, since it was not Malaysian, can be assume at oversea level. It's ok better than nothing right ...


Friday, August 7, 2009

Plan to do attachment program at Singapore Univ.

As student sponsor by NSF, I was allowed to do attachment at any other University to get more experience and increase learning curve during studies. For master student we are allow only at Malaysia and Singapore Univ. only while for PhD student unlimited to any Univ. in the world. I do some finding on University at Singapore. There are NUS and NTU. So I choose NTU because this university more focus on technology area compare with NUS.

One lecture ware interested to supervise me during attachment program, he was Asst Prof Low Yoke Hean Malcolm. His area of specialization on optimization of traveling salesman problem, job shop scheduling and resource scheduling. Hope fully everything run smoothly ...

Sunday, July 26, 2009

ICSECS'09 Camera ready secure uploaded


Yesterday, since there is no negative comment from reviewer so I just upload my paper direct to website. Now wait for college to approval sponsor for this conference. That all ...

Thursday, July 2, 2009

ICSECS '09 conference accepted

Saya dengan rasa rendah diri mengucapkan kesyukuran pada Allah daku sangat kerdil terhadap kudrat Mu ya Allah ...

From: icsecs09 [mailto:icsecs09@ump.edu.my]
Sent: 02 Julai 2009 16:52
To: Mohd Taufiq Norhizat
Subject: ICSECS’09 : Acceptance Notification for paper ID 252

Dear Prof./Dr./Mr./Ms. Mohd.Taufiq Norhizat, Thank you for your submission to International Conference of Software Engineering & Computer Systems 2009 (ICSECS’09). Congratulations, on behalf of the Mohd.Taufiq Norhizat, we are pleased to inform you that your submission,
----------------------------------------------------------------------
Paper ID: 252
Title: Study of performance comparison of AODV and DSDV routing protocol algorithm with difference number of nodes for wireless sensor network
----------------------------------------------------------------------
has been conditionally accepted for oral presentation in this conference. We have included the reviewers' comments at the end of this message. For further query, please do not hesitate to contact us. We are looking forward to meeting you at Malaysia, the truly Asia.
Best Regards,
Program Committee, ICSECS’09
---------------------------------------------------------------------
Comments to Author:
paper highlights a network issues with one of the focuses parameter is delay of data transmission that relate to the conference area. The contents of the paper are informative to get basic principles. Good for reference.

Friday, June 26, 2009

ICOCI 09 conference at The Legend Hotel KL

This week the ICOCI 09 conference was held at The Legend Hotel KL and it was my 3'rd conference that I was attended. During my presentation one of the audience was so impressed with my project. It was good opportunity to meet with other researchers from other university. All of them presented with good presentation. Then I also join other session to learn some thing new about new technique especially some thing that related with Genetic Algorithm which that I want to use it for my next algorithm.
.

Thursday, June 11, 2009

Chairman for conference session

Salam every one,
I received email wrote:

Assalamualaikum,

Sukacita dimaklumkan bahawa tuan telah dilantik sebagai Pengerusi Pembentangan Sesi Selari ICOCI09. Untuk makluman, pengerusi dilantik di kalangan pembentang dalam sesuatu sesi. Maklumat lanjut pembentangan adalah seperti berikut:

Tarikh : 25 Jun 2009 (Khamis)
Masa : 10.40 - 12.30pm
Sesi : 15

Diharap tuan dapat membalas emel ini selewat-lewatnya pada 9 Jun 2009 bagi menyatakan penerimaan/penolakan perlantikan tersebut. Surat perlantikan rasmi akan diserahkan pada hari pendaftaran ICOCI09.

Sekian, terima kasih.

Dr Yuhanis Yusof
AJK ICOCI09
Universiti Utara Malaysia


Is great pleasure for me to experience this opportunity ...

Tuesday, May 19, 2009

ICSECS'09 paper conference secure uploaded

Yesterday after having reviewed by Madam Azlina for these paper, she make correction to a few mistake that I have made. The paper secure uploaded to respective conference website. I hope this paper can be accepted without any problem. As I know Faridzul groups also will be submitting one paper here. I also hope that I can go there to present my paper together with other group. It will be held at Spa Swiss Garden Resort, Kuantan, it was just lovely place to relax after struggle to complete this paper. Enjoy ...

Friday, May 15, 2009

ICOCI 09 paper sponsored ready

Last week I receive mail wrote:


Dear Sir/Madam
We refer to the above matter. We are pleased to inform you that Universiti Tenaga Nasional, Malaysia will be sending 1) Mr. Mohd Taufiq Norhizat to attend the above conference. We would appreciate if you send us your confirmation of our registration for our further action. Kindly please forward us the Invoice for the payment as soon as possible. We will process after receiving the original invoice.
Any inquiries regarding the above matter, you may contact:

Telephone No. : (603) 89212020 ext 1252/1257
(Attn: Ms Nor Azira/ Ms Raihana)

It show me that my paper was sponsored by university for respective conference which it will be held at The Legend Hotel KL in June, 24 to 25 2009. This is the great opportunity for me and then I can add one more conference proceeding for my collection. For your information I was one of presenter in this conference as my third paper conference in hand and also not forgotten along together with me at there was 4 people from my group project come as normal participant some sort of "pemerhati" for our work. Just for now what I need to prepare is my conference slide presentation.
That all thank you

Thursday, April 23, 2009

My paper now citated at IEEE Explore

Just now while I checking at IEEE Explore then I found that my paper already in IEEE citation. Alhamdulilah, what a beautiful day ...
If you want to see my paper you can just go to IEEE website at http://ieeexplore.ieee.org/ then type searching for title "VoIP over MYREN" after that you can easily see my paper ...

Sunday, April 12, 2009

How to fix DYMOUM trace file can be use with tracegraph

Last week one final year student see me asking about problem how to analyze DYMO trace file using tracegraph, then I try to investigate what is the problem. When open DYMO trace file using tracegraph, it shown some error that about some sort "ip bla.. bla.. thing". That make me to do a comparison between other routing protocol trace file with DYMO trace file. Then I found that there is some new element that tracegraph cannot understand. So the easy way to solve it, just to delete that "not understand" data.
  1. By using Notepad++ software open that trace file.
  2. Then open find & replace menu.
  3. Select regular expression search mode. See figure attach
  4. Then at find what: type \[ pt . \]
  5. For replace with: leave it empty
  6. Replace all
  7. Dont forget to save your file

Your trace file now can open using tracegraph

Ok, peace

Thursday, April 9, 2009

ICOCI09 conference paper accepted

This morning I receive letter wrote:

Dear Sir / Madam,

We a re pleased to inform you that your paper has been accepted with minor modification for oral presentation at the International Conference on Computing and Informatics (ICOCI09).
Please check the reviewer’s comment in the attachment. The format for camera-ready can be obtained from the conference website under the ‘Submission’ page. The camera-ready file must be submitted to us before or on 14th April 2009.

Please be informed that without the camera ready version, there is no guarantee that your paper will be included in the proceedings. Further details on registration and payment can be obtained from ICOCI09’s website. Kindly return your camera-ready manuscript to shafinah@uum.edu.my.

Thank you.

Regards,
Shafinah Farvin Bt. Packeer Mohamed
Committee Member of Conference Paper and Proceedings
ICOCI09

Ya Allah sesungguh nya pemberian ini sangat berharga buat ku ...

Tuesday, February 17, 2009

New swarm intelligent algorithm will be implement

After attend talk by Prof. Abbas Jamalipour at UPM today then I got inspire from his speech. It was about new algorithm on swarm intelligent technique to solve network problem. One of the method is how Bee worker communicate with each other and how their distribute their job like queen bee, worker bee, soldier bee & etc. He also tell about another method which is ant algorithm & how the way they searching path or route to their food. So I will try to propose implement his idea to my new algorithm, perhaps it can solve some issue...

New algorithm give a lot off effect on dropped packet

After I satisfy for extended life time for battery of WSN router. It also need to look up on dropped packet using trace graph. Then I realize that when certain node is push to sleep mode if I cannot reroute properly than a lot of dropped packet will happen. He I show that quite a lot of dropped packet compare than other routing protocol.

Saturday, February 14, 2009

New algorithm with more than 3000 sec.

I try new algorithm with more than 3000 sec. it was suggested by Assoc. Prof. Dr. Norashidah. Then suddenly I saw that my algorithm is not good enough to save energy for long term utilization. Then I do some modification that to be fair for all intermediate nodes. So I come out with new method where every intermediate node will have same time of sleep. To calculate that we record when 1'st node come to sleep then until the last node start to sleep. So the same duration will apply to other intermediate nodes.
Some sort like this:

∑ sleep time = Time last node sleep - Time 1'st node sleep
Time node i wake up = Time node i sleep + ∑ sleep time

Then the result change where new method can save more energy compare with other routing protocol.


Tuesday, February 3, 2009

AWK file solution for JTrana

Last mid night while working with Zul, I just only thinking how to solve problem of wasting time waiting JTrana process. The problem is when we want to show energy graph it take about half to 1 day for analyzing trace file. Like example when I set simulation time until 3000 second then size the trace file increase up to 0.5 Gbyte produce long time to analyze. Then our greatest sifu Zul come to introduce how to use AWK file. Then after few basic step he teach me, I realize that with AWK script file can solve problem waiting from 1 day become few second. I so wonderful, how most great Allah introduce someone like him to me.

Saturday, January 31, 2009

Library modular define in C short course

Last week I attend short course teached by our sifu Madam Marina, she was one of C programming sifu in CoIT. It was very nice to know how library modular define work. One long program can be divide by 3 main file:
  1. Header file (for global variable, function declaration, constant)
  2. Application file (for main body)
  3. Implementation file (for function definition)
Each of this file have their own function to make the program more easier. If you interested to learn you can log in Here


Tuesday, January 27, 2009

Dropped packet for new routing strategy but still maintain same as existing RP

Again I look on dropped packet for new energy strategy give same amount dropped packet with DSDV routing algorithm. So here I can conclude that our new strategy is still not bad. This is 1'st our objective achievement where we concentrating improvement on energy. The result:

Wednesday, January 21, 2009

New energy routing strategy give side effect on packet delay

After I try new energy routing strategy on my simulation environment that can give advantage on remaining energy. But then when I look at packet delay show that new strategy effected. It is because after one or two intermediate nodes in sleep mode where data from source node have to go through long way until reach access point. Here the result:


We can see 3 out of 5 nodes the new algorithm is better which is node1, node 2 and node 7 but two more nodes effected cause of new energy routing strategy. So now I need to fine another way that can improve in this problem...

Saturday, January 17, 2009

Energy saving strategy worked ... !

After I start testing our energy strategy on simulation environment. We can see from 1'st result that show some improvement on energy saving on remaining energy. Then after compare with original simulation without energy strategy, our claim that it can prolong energy life time of battery.



Further more there is need to check whether this energy strategy can ensure network performance always in good condition...


Sunday, January 11, 2009

Flow chart new method

Today I want to introduce new method in energy saving strategy. It was focus on intermediate node, where each node rotating shift sleep mode due to prolong remaining energy. Here I show where we got 2 level of energy threshold value where 70% and 35% of remaining energy. Any intermediate node will change to sleep mode when 70% of remaining energy or level 1 threshold value accord. But if all intermediates node exceed level 1 threshold value, the threshold value itself will change to level 2 threshold value where until 35% of remaining energy. Then the same process apply to intermediate node.

Wednesday, January 7, 2009

Set node to idle/sleep state

To further our routing algorithm, there is need to do some saving energy where the specific node can be force to sleep or idle modes. Our focus is in intermediate node where the node will always exhaust energy because of forwarding data from source node to destination node. But working with NS-2 with physical layer mac 802.11 there is no basic command that can set node to be sleep/idle. The node itself can sleep/idle when there no activity sending or receive data on node. So there is need to figure out some alternative way to make the node force to sleep.

Newer Posts Older Posts Home

 

blogger templates | Make Money Online