From CSWiki

To enter a new comment:

• Copy the following template.

==Your subject. ==
Your comment.

—~~~~

• Click [edit] to the right below the following line.
• Past the copied template immediately below the first line of the edit box (= =).
• Substitute your content for Subject and Your comment, leaving the rest as it is.
• Click the Save page button.

[edit]

[edit] Nine non-technical skills

Note: Nine important skills.

As far as this class is concerned, note especially 6, 7, and 9.

Russ Abbott 10:23, 20 November 2007 (PST)

[edit] Proposed BA program in Information Technology

The Department is developing a BA program in Information Technology. As part of the vetting process we are asking interested parties (including our current students) to take a very brief survey regarding their reaction to the proposal.

The survey describes the differences between the current BS program and the proposed BA program. It then asks the following questions.

1. Your name (Optional):

2. Your position: (student, alumnus, etc.)

3. From your perspective, is the BA IT a worthwhile undergraduate option within the 
   Department of Computer Science?

   Yes

   No

4. From your perspective, which program would you prefer?

   BS CS

   BA IT

5. Please elaborate on your choices for #3 and #4, and leave any additional comments
   you would like to share with us: (Optional)

Please take the survey and tell us what you think.

Thanks.

Russ Abbott 09:57, 13 November 2007 (PST)

[edit] Note: JPL Student Internship program

This is a good program. You should consider it. The first step is to contact Christ Khachikian at Christ Khachikian

[edit] Heap space and VM arguments

I include the following flags as VM arguments in eclipse. -ea -Xmx900m

The first turns on assertions. The second asks for 900MB of heap space. Pick a number that your computer can support. Also adjust BoatTorpedoModel.seaSide to fit your computer. It is currently set to 800. A smaller number requires less memory, but it makes the boat and torpedo images smaller.

Russ Abbott 07:53, 10 November 2007 (PST)

[edit] New version uploaded

Another new version has been uploded. We'll talk about it in class Saturday. To be informed when a new version is uploaded, put a watch on the upload page.

Russ Abbott 16:14, 26 September 2007 (PDT)

---

I've uploaded a new version of the Boat-Torpedo code.

• Right after class Aaron told me about a bug fix which makes the leggy tree search more effective.
• This evening I made a couple of additional minor changes. The most visible draws a border around the space and reverses the torpedo colors. Torpedoes are now white unless they are seen as dangerous at which time they are drawn as red.

—Russ Abbott

[edit] Sat 29 Sept, 2007 Chapter 3

I have read chapter 3 and I could present the material to the class if no one objects.

-Richard

[edit] Sat 6 th Oct, 2007 Chapter -3

Today discussed about boat-torpedo problem. First I am going to get the thoritical knowledge about searching technique.So,I decide to stay in a theoretical part.So,I am going to prepare for the presentation of 3 rd chapter. Richard is giving presentation on this week so next week I am going to prepare for presentation.Actually I tried to read the slide but i can not understand some of the topic.So,I am going to buy or find the book from library and prepare for it.I would alos like to be in 6ht chapter too.

Kunjal

[edit] Chapter 6 (Sept 24)

Like Kunjal, I would also like to stay in the theoritical part of this project. Since there is already 2 people working on the presentation for chapter 3 I will look over chapter 6 and develope a presentation.

-Michael

[edit] Subversion Accounts

Thanks for the comments on Subversion. I've deleted the discussion (although it is available in the archived history pages) since we now have Subversion accounts.

[edit] Armando Padilla position.

I've aways wanted to be in a R&D project team so I would like to stick with the theoretical and maybe introduce a new search algorithm into the application. I will look into other possible search algorithms that can be applied and go from there.

Also If anyone needs a hand with either Chapter 6 or setting up SVN (subversion) let me know. Ill be more than happy to give a hand on that aswell. I went ahead and created the below table to give the positions a bit more structure so we can see what everyone will be working on.

[edit] CS460 Positions (put you name into the desired box. Grouping based on the groups jotted down on board)

I've moved this list the the main course page. Please sign up there.

Thanks Armando for setting it up.

[edit] Armando's Comment :: Torpedo strategies

I was looking at the sections on the main page, specifically the section on allowing torpedoes to work together. I was wondering if maybe someone should look into the different strategies of pack hunting animals such as lions, wolfs, and maybe raptors? i was thinking along the lines of having one torpedo fool the boat into moving one way so the other torpedoes could kill the boat much like lions do in the wild.

- Armando

If I understand the situation correctly, the boat needs to know exactly which torpedo would go where when it calculate "how good" is any of the path it generates at every time step. That means it should not be possible to "fool" a boat, unless the class/professor agree to change the premise of this simulation. --James Cheng 14:55, 29 September 2007 (PDT)

Yes, the challenge is that the torpedoes must be completely predictable so that the boat can determine how they will respond to its moves. But the boat is not predictable to the torpedoes. They can't ask how the boat will respond to its moves. All the torpedoes know are the boat's generic limitations such as its turning and speed ranges. Russ Abbott 23:19, 30 September 2007 (PDT)

[edit] Code consistency

I noticed while coding that the code uses 2 spaces instead of tabs. I think it would be a good idea for everyone to set up eclipse so that it doesn't insert extra tabs. Preferences->Java->Code_Style->Formatter->Edit...(Tab Policy: Spaces Only ; Indentation Size: 2)

If anyone wants to set other guidelines on how the code should look, say so; we could also discuss it in class on Saturday.

[edit] Netbean IDE for the Project

I am a fan of Netbean IDE. Since I feel Netbean's UI is more intuitive than Eclipse's, I decided to use Netbean IDE for my project. Netbean 6.0 has a versioning module which supports SVN natively, so you could use SVN without any plugin or additional setup. Although currently only beta version is available, the stable version will be out around end of November. So if you are interested, it still worth to give it a try.

More info: http://www.netbeans.org/community/releases/60/index.html

—Ctham 01:29, 1 October 2007 (PDT)

[edit] Initial Position

At the begining of the simulation, completely random positions of boat and torpedos occasionally makes the boat be caught or reach home station right away. Why don't we set the initial position of boat have to be away from the home station at least curtain distance and positions of torpedos have to be away from the boat as well. This would bring more accurate statistics. --Kunhan Kim 16:04, 4 October 2007 (PDT)

Good idea. It's already done. The current version has fixed positions for the boat and home base. The torpedoes are placed randomly but not in the same quadrant as the boat. Russ Abbott 16:06, 4 October 2007 (PDT)

[edit] Profiling

Professor mentioned profiling the system to see how it utilize the CPU resource; here's Extensible Java Profiler, which may be helpful. --James Cheng 14:34, 6 October 2007 (PDT)

[edit] Torpedo speed and numbers of torpedoes

The question of how many torpedoes you need to catch the boat suggested an experiment. With the TorpedoSpeedMultiplier moved up from the default of 2.5 to 3.0, a single torpedo almost always catches the boat, using either search technique. So a value of about 2.5 seems to be close to a critical value. Above that, a single torpedo is too much for the boat. At that value and below, the boat is pretty good at escaping from 5 torpedoes. At 2.8 the boat seems to manage to escape more often than not — but it's often close.

Is there any obvious mathematical reason why a speed factor of about that size is such a sensitive parameter?

Recall that the TorpedoSpeedMultiplier is the ratio of the torpedo to the boat. The torpedo speed is set to be that factor times the boat speed.

Russ Abbott 20:00, 6 October 2007 (PDT)

[edit] Re: Torpedo speed and numbers of torpedoes

Here's a mathematical way to frame the question. If you put a single torpedo at the base headed away from the boat — so that it has to go "around the world" the long way as the fastest way to get to the boat — at what speed ratio will it always catch the boat? Of course that depends on their relative turn ratios as well. If the boat starts turning immediately, will it be able to get out of the torpedo's range in time? Can you come up with an analysis that characterizes this simplified problem?

[edit] Subversion Current Status

Still subversion server cs1 cannot be access from off campus. Campus firewall is blocking the port and cs department had requested to open the port. This process is taking longer than I thought. I have an alternate solution--using my server for the time being until the port problem gets resolved. But, I would like to ask Dr. Abbott for the alternative solution. Please visit Subversion (Subclipse) for complete subversion, subclipse and accounts information.

-- Manoj Katwal 08:51, 13 October 2007 (PDT)

[edit] Executable Jar File

When running executable jar file created by Eclipse, exception is thrown. I wrote a batch file to correct the classpath. Update repastPath and mainJarFile accordingly, the jar file should run.

 set repastPath=D:\My Document\java\cs460\BoatTorpedoes\dist\lib
 set mainJarFile=BoatTorpedoes.jar

Note that repastPath should contain repast.jar and all the jar files under the lib directory of repast distribution.

Media:Run.zip

--Ctham 23:22, 15 November 2007 (PST)

Thanks. But the real value of a jar file is so that one can send a double-click-executable file to someone else. Russ Abbott 23:39, 15 November 2007 (PST)