Scholarlink

<p>Web design primarily revolves around fixed templates and static user flows, providing the best limited responsiveness to user context or behavior. With the digital world becoming increasingly sophisticated, there is an increasing demand for websites that respond in real-time to user inputs, preferences, and behavior patterns. Thus, the given paper introduces a hybrid framework that utilizes the generative capability of large language models (LLMs) and the optimization effectiveness of metaheuristic algorithms. The proposed framework comprises six components, namely, context analyzer (CA), prompt technique, LLM content generator, evaluation engine, metaheuristic optimizer, and rendering loop. LLMs generate and transform website components in real-time. By integrating this generation ability with a genetic algorithm (GA), we facilitate dynamic website layout optimization, content, and user interface optimization. A cost function is also proposed that measures important performance metrics like load time, bounce rate, and conversion rate and demonstrates how metaheuristic optimizers such as genetic algorithms (GAs) can optimize this cost in various scenarios. The proposed framework greatly enhances user interaction, responsiveness, and website performance, thereby leading to automated optimization.</p><p>.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

LLM-GA: A hybrid framework to build dynamic websites for optimizing web performance

  • 期刊论文

  • 发表日期：

    2025年8月11日

查看全文
  • Bhanu Sekhar Guttikonda,
  • Ram Chandra Sachan,
  • Vinod Veeramachaneni

摘要

Web design primarily revolves around fixed templates and static user flows, providing the best limited responsiveness to user context or behavior. With the digital world becoming increasingly sophisticated, there is an increasing demand for websites that respond in real-time to user inputs, preferences, and behavior patterns. Thus, the given paper introduces a hybrid framework that utilizes the generative capability of large language models (LLMs) and the optimization effectiveness of metaheuristic algorithms. The proposed framework comprises six components, namely, context analyzer (CA), prompt technique, LLM content generator, evaluation engine, metaheuristic optimizer, and rendering loop. LLMs generate and transform website components in real-time. By integrating this generation ability with a genetic algorithm (GA), we facilitate dynamic website layout optimization, content, and user interface optimization. A cost function is also proposed that measures important performance metrics like load time, bounce rate, and conversion rate and demonstrates how metaheuristic optimizers such as genetic algorithms (GAs) can optimize this cost in various scenarios. The proposed framework greatly enhances user interaction, responsiveness, and website performance, thereby leading to automated optimization.

.