Adam Viktorin

This paper presents EASE (Effortless Algorithmic Solution Evolution), an open-source and fully modular framework for iterative algorithmic solution generation leveraging large language models (LLMs). EASE integrates generation, testing, analysis, and evaluation into a reproducible feedback loop, giving users full control over error handling, analysis, and quality assessment. Its architecture supports the orchestration of multiple LLMs in complementary roles - such as generator, analyst, and evaluator. By abstracting the complexity of prompt design and model management, EASE provides a transparent and extensible platform for researchers and practitioners to co-design algorithms and other generative solutions across diverse domains.

This paper presents the competition entry on LLM-designed evolutionary algorithms for the GNBG-generated benchmark suite. We use EASE (Effortless Algorithmic Solution Evolution), a modular framework that prompts a large language model to generate complete optimizers, evaluates them under the competition protocol, and feeds performance summaries together with improvement-oriented analysis back into subsequent iterations. The automated process produced ten valid candidate algorithms. The best generated solver was a hybrid adaptive Differential Evolution method combining L-SHADE-style parameter adaptation, archive-based diversity preservation, covariance-inspired variation, and budget-aware local refinement. The results show that iterative LLM-guided refinement can produce effective, benchmark-specific optimizers.

This paper provides a detailed analysis of the potential technical and architectural challenges associated with building the FrontEASE software. FrontEASE serves as the web wrapper, utility extension, and user interaction layer for the modular LLM-based optimization framework, EASE (Effortless Algorithmic Solution Evolution). The focus is on the architectural choices made and how this unconventional solution, designed to streamline the use of LLM-based optimization workflows for scientific applications, aligns with best practices in standard software engineering. The primary goal of this paper is to promote and facilitate further experimentation with enterprise-grade architectures for LLM-driven applications. Ultimately, this work aims to guide the development of these products into reliable and maintainable software solutions over the long term.

Modern business applications increasingly demand chatbots that can perform complex tasks by augmenting large language models (LLMs) with external tools and data sources. In this paper, we propose a modular and extensible chatbot architecture that integrates LLMs with tools such as relational databases, semantic search engines, and computational APIs to enhance the bot’s capabilities. A key novelty of our work is a comprehensive benchmarking methodology designed to systematically evaluate the performance of these tool-augmented chatbots across diverse tasks and scenarios. We implement the architecture using contemporary LLM technology and evaluated it on realistic business queries requiring multi-step reasoning. Experimental results demonstrate that the augmented agent can successfully handle complex queries that demand extended context, precise computation, and domain-specific knowledge-capabilities that remain challenging for a standalone LLM. The proposed architecture and guidelines are model-agnostic and future-proof: it allows new tools or more advanced LLMs to be plugged in with minimal effort, and provides a reproducible evaluation suite to measure their impact.