Case-Based Genetic Optimization Of Web User Interfaces

The combination of case-based approach and genetic optimization can provide significant boost to effectiveness of computer-aided design of web user interfaces (WUIs). However, their integration in web design domain requires certain sophistication, since parts of available solutions cannot be reused directly, due to technical and legal obstacles. This article describes evolutionary algorithm for automatic generation of website designs, which treats parameters of functionality, layout and visual appearance as the variables. The structure of the chromosome is devised, allowing representation of websites' properties in the above three manipulated aspects and facilitating easy application of the genetic operators. We also describe organization and population of repository of filler-up content, which is compulsory for evaluation of WUI fitness with regard to the needs and preferences of users. We demonstrate retrieval of web designs as cases and propose using similarity measure in the fitness function to adapt the generated WUI to these examples. Finally, implementation of the approach is illustrated based on the popular Drupal web framework. The results of the study can empower case-based reuse of existing web designs and therefore be of interest to both AI researchers and software engineers.