| Automatic Code Generation (ACG) allows software engineers to create more concise, maintainable, and reusable solutions, ultimately improving their productivity. Despite the significant benefits and the profound economic impact of ACG in the software development field, it still often requires substantial input and interaction from humans. The presented Automatic Coder using Artificial Intelligence (ACAI) system uses a novel approach to solve fully automated code generation in routine programming domains. Given high level user goals and preferences, a library of abstract programs, and a library of generic code components, ACAI uses a combination of Case-Based Reasoning, Routine Design, and Template-Based Programming approaches to generate complete Java programs that satisfy user requirements. |
