Publication: Ejecución y adaptación de trazas de juegos para la automatización de pruebas
Loading...
Official URL
Full text at PDC
Publication Date
2015
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
Tradicionalmente, las empresas de videojuegos y los desarrolladores invierten mucho tiempo y recursos en probar que los videojuegos siguen teniendo el comportamiento esperado después de realizar ciertos cambios en él. Este trabajo lo llevan a cabo lo que conocemos como testeadores humanos, que juegan una y otra vez los mismos juegos y niveles con el objetivo de detectar fallos. Debido a la naturaleza de los videojuegos, la automatización de las pruebas mediante soluciones como tests de unidad es prácticamente imposible.
En este trabajo se propone una forma alternativa de realizar tests, que se basa en la reproducción de trazas adaptadas que se han generado y guardado anteriormente al jugar el testeador humano. Con este método, si los cambios en el juego son mínimos, si son cambios a nivel de programación, etcétera, se podrá comprobar que su comportamiento es correcto sin ningún esfuerzo adicional.
Para llevar a cabo nuestro objetivo, proponemos un sistema externo de generación, salvado y reproducción de trazas, así como también un método de adaptación de las mismas que utiliza el razonamiento basado en casos y la inteligencia artificial ya existente en el videojuego usado como base (Blázquez Checa et al., 2013-2014) para llevar a cabo las acciones necesarias para completar los objetivos sin ningún error.
Traditionally, game industry companies and developers spend a lot of time and resources in testing videogames to prove they keep having the same expected behaviour after performing some changes at it. This job is carried out by what we know by human testers, who play the same gamesand levels time and time again with the purpose of detecting errors. Because of the nature of videogames, automation of testing with solutions like unit tests is almost impossible. Here we propose an alternate form of making and passing tests, based on the reproduction of adjusted traces that where generated and saved before, when the human tester was playing the game. With this method, if the changes in the game are minimum or if they are programming-only changes then checking that the behaviour is correct can be done without any additional effort. In order to complete our goal, we propose an external system that generates, saves and reproduces traces, as well as a method for adapting these traces. This method uses case-based reasoning and the artificial intelligence that previously existed in the game used (Blázquez Checa et al., 2013-2014) to accomplish the actions needed to complete all the objectives without making any mistakes
Traditionally, game industry companies and developers spend a lot of time and resources in testing videogames to prove they keep having the same expected behaviour after performing some changes at it. This job is carried out by what we know by human testers, who play the same gamesand levels time and time again with the purpose of detecting errors. Because of the nature of videogames, automation of testing with solutions like unit tests is almost impossible. Here we propose an alternate form of making and passing tests, based on the reproduction of adjusted traces that where generated and saved before, when the human tester was playing the game. With this method, if the changes in the game are minimum or if they are programming-only changes then checking that the behaviour is correct can be done without any additional effort. In order to complete our goal, we propose an external system that generates, saves and reproduces traces, as well as a method for adapting these traces. This method uses case-based reasoning and the artificial intelligence that previously existed in the game used (Blázquez Checa et al., 2013-2014) to accomplish the actions needed to complete all the objectives without making any mistakes
Description
Trabajo de Fin de Grado en Ingeniería Informática (Universidad Complutense, Facultad de Informática, curso 2014/2015)