| Both sides previous revisionPrevious revisionNext revision | Previous revision |
| public:t-622-arti-09-1:lab_1_materials [2009/01/19 14:23] – arnarbi | public:t-622-arti-09-1:lab_1_materials [2024/04/29 13:33] (current) – external edit 127.0.0.1 |
|---|
| === Lab 1 material: Lisp === | === Lab 1 material: Introduction to Lisp === |
| |
| * [[http://cs.gmu.edu/~sean/lisp/LispTutorial.html|Lisp Tutorial]] | * [[http://cs.gmu.edu/~sean/lisp/LispTutorial.html|Lisp Tutorial]] |
| * [[http://www.gigamonkeys.com/book/lispbox/|LispBox - easy one-install Lisp implementations]] | * [[http://www.gigamonkeys.com/book/lispbox/|LispBox - easy one-install Lisp implementations]], Locally:[[http://www.ru.is/faculty/hannes/share/lispbox07.exe|lispbox07.exe]] [[http://www.ru.is/faculty/hannes/share/Lispbox-0.7-with-acl81_express.dmg|Lispbox-0.7-with-acl81_express.dmg (OS X 10.4 PPC)]] |
| * [[http://www.apl.jhu.edu/~hall/Lisp-Notes/Introductory-Lisp-Overview.ps|Lisp cheat sheet (PS)]] | * [[http://www.apl.jhu.edu/~hall/Lisp-Notes/Introductory-Lisp-Overview.ps|Lisp cheat sheet (PS)]] {{:public:t-622-arti-09-1:introductory-lisp-overview.pdf|(PDF)}} |
| * More material for learning Lisp: | * More material for learning Lisp: |
| * [[http://mypage.iu.edu/~colallen/lp/|Lisp Primer]] - A tutorial | * [[http://mypage.iu.edu/~colallen/lp/|Lisp Primer]] - A tutorial |
| * [[http://refcards.com/docs/wingb/xemacs/xemacs-refcard-a4.pdf|Emacs reference card (PDF)]] | * [[http://refcards.com/docs/wingb/xemacs/xemacs-refcard-a4.pdf|Emacs reference card (PDF)]] |
| |
| Work through the following exercises. Refer to the above material for more information. | To work through these exercises, refer to the above material for more information. |
| |
| - **Emacs and REPL:**\\ Install [[http://www.gigamonkeys.com/book/lispbox/|LispBox]] on your computer and start it up. The lower window is the Lisp REPL (read-eval-print loop) where you can enter Lisp expressions and have them evaluated. | - **Emacs and REPL:**\\ Install [[http://www.gigamonkeys.com/book/lispbox/|LispBox]] on your computer and start it up. The lower window is the Lisp REPL (read-eval-print loop) where you can enter Lisp expressions and have them evaluated. |
| * In the REPL, type '','' (comma); this invokes the command mode of the REPL, type the command ''cd'' and press enter. Emacs will prompt you for a file path, enter the file path to the folder where you saved the file. | * In the REPL, type '','' (comma); this invokes the command mode of the REPL, type the command ''cd'' and press enter. Emacs will prompt you for a file path, enter the file path to the folder where you saved the file. |
| * Now type (in the REPL) the expression ''(load "agents.lisp")''. | * Now type (in the REPL) the expression ''(load "agents.lisp")''. |
| * Now the definitions of the file have been loaded in the REPL. To see what definitions are available, open the file in the upper half of Emacs. To do this, you can click in the upper half and then press (in sequence) Cmd+C and Cmd+F and type the path to the file. Alternatively just select "File -> Open" :-) If Emacs asks you to allow it to apply local variables from the file, then answer with Yes. | * Now the definitions of the file have been loaded in the REPL. To see what definitions are available, open the file in the upper half of Emacs. To do this, you can click in the upper half and then press (in sequence) Ctrl+X and Ctrl+F and type the path to the file. Alternatively just select "File -> Open" :-) If Emacs asks you to allow it to apply local variables from the file, then answer with Yes. |
| * Have a look at the file and see if you can understand the code. Don't worry too much if you don't. At the bottom of the file there are three example expressions. Try typing (in order) them in the REPL and see what happens | * Have a look at the file and see if you can understand the code. Don't worry too much if you don't. At the bottom of the file there are three example expressions. Try typing (in order) them in the REPL and see what happens |
| - **Writing an agent program:** \\ The function ''EXAMPLE-AGENT-PROGRAM'' gives an example of how to create a program to control an agent in the vacuum-cleaner world. The program is a function that takes a keyword list representing the percept value, and it should return one of the symbols ''LEFT, RIGHT, UP, DOWN, SUCK'' or ''IDLE''. A percept keyword list looks like this: <code lisp> | - **Writing an agent program:** \\ The function ''EXAMPLE-AGENT-PROGRAM'' gives an example of how to create a program to control an agent in the vacuum-cleaner world. The program is a function that takes a keyword list representing the percept value, and it should return one of the symbols ''LEFT, RIGHT, UP, DOWN, SUCK'' or ''IDLE''. A percept keyword list looks like this: <code lisp> |
| '(:x 3 :y 8 :dirty T) | '(:x 3 :y 8 :dirty T) |
| </code> \\ Here, '':x'' and '':y'' give the location of the agent and '':dirty'' tells us if the current square has dirt or not. The given example program is an implementation of the code in figure 2.8 on page 46 in the textbook. \\ Write a new agent program, that walks about the environment randomly, sucking up dirt where it finds it. You can either write your function in the REPL or add it to the file. If you add it to the file, you have to save the file (keyboard shortcut is Ctrl+X, Ctrl+S) and then reload the file in the REPL (recommended option). Instead of typing the load command again, you can reload the file by hitting Ctrl+C, Ctrl+L while the file window is active. \\ Use the ''SIMULATE'' and ''SIMULATE-QUIET'' functions from last step to test your new agent program. Does it give better performance evaluation than the old one? \\ | </code> Here, '':x'' and '':y'' give the location of the agent and '':dirty'' tells us if the current square has dirt or not. The given example program is an implementation of the code in figure 2.8 on page 46 in the textbook. \\ Write a new agent program, that walks about the environment randomly, sucking up dirt where it finds it. You can either write your function in the REPL or add it to the file. If you add it to the file, you have to save the file (keyboard shortcut is Ctrl+X, Ctrl+S) and then reload the file in the REPL (recommended option). Instead of typing the load command again, you can reload the file by hitting Ctrl+C, Ctrl+L while the file window is active. \\ Use the ''SIMULATE'' and ''SIMULATE-QUIET'' functions from last step to test your new agent program. Does it give better performance evaluation than the old one? \\ |
| - **Writing a performance evaluation function:** \\ The function ''EXAMPLE-MEASURE'' shows how performance evaluation functions look. This function takes an environment and an agent as parameters, and returns a number indicating the performance of the agent so far. \\ Both the agent and the environment are so called //structures//. Their fields are defined in the file ''agents.lisp'' with calls to the ''DEFSTRUCT'' macro. This macro defines a function for constructing an instance of the structure, e.g. for the ''ENV'' structure, this function is called ''MAKE-ENV''. \\ The macro also defines functions for accessing the fields of a structure, e.g. the //width// field of an ''ENV'' structure in variable //env// can be accessed with ''(env-width env)''. \\ Write a performance evaluation function that awards 10 points for every clean square but deducts one point for every move the agent makes. The number of moves is stored in the //moves-so-far// field. Now run the simulations again. How does this change the evaluation score? Can you think of ways to change the agent program so that it scores better with this new evaluation function? | - **Writing a performance evaluation function:** \\ The function ''EXAMPLE-MEASURE'' shows how performance evaluation functions look. This function takes an environment and an agent as parameters, and returns a number indicating the performance of the agent so far. \\ Both the agent and the environment are so called //structures//. Their fields are defined in the file ''agents.lisp'' with calls to the ''DEFSTRUCT'' macro. This macro defines a function for constructing an instance of the structure, e.g. for the ''ENV'' structure, this function is called ''MAKE-ENV''. \\ The macro also defines functions for accessing the fields of a structure, e.g. the //width// field of an ''ENV'' structure in variable //env// can be accessed with ''(env-width env)''. \\ Write a performance evaluation function that awards 10 points for every clean square but deducts one point for every move the agent makes. The number of moves is stored in the //moves-so-far// field. Now run the simulations again. How does this change the evaluation score? Can you think of ways to change the agent program so that it scores better with this new evaluation function? |
