A Generator for Creating Streaming Continuous Optimisation Benchmarks - Supplementary Materials

Le résumé fourni par la source

SCOBench: Streaming Continuous Optimisation Benchmark Nemeth, M.B., Hart, E., Sim, K., Renau, Q. (2027). A Generator for Creating Streaming Continuous Optimisation Benchmarks. In: Iacca, G., et al. Parallel Problem Solving from Nature – PPSN XIX. PPSN 2026. Lecture Notes in Computer Science, vol 16985. Springer, Cham. https://doi.org/10.1007/978-3-032-36214-8_30 This repository contains source code for the benchmark, along with supplementary materials. SCOBench is a fully configurable generator that can generate streams of instances with different drift patterns specified by the user. In contrast to existing stream generators that define drift as a shift in the location of the optima over time of a single instance, SCOBench generates a stream of new instances from multiple functions, while drift between consecutive instances can be defined according to a change in three different metrics (landscape features, probing-trajectories and optima location). Use of the benchmark: The code contains a simple example file showcasing stream generation. All drift, similarity, and search parameters can be customised using the benchmark config class. Parameters for the stream include problem dimension, BBOB function(s), and drift pattern (drift types supported: 0-no drift, 1-major drift, 2-re-major drift, 3-minor drift, 4-re-minor drift). Using the generate_problem_stream function, an ordered set of optimisation problems is generated using IOH corresponding to the provided config and parameters; additionally, all generation details are provided in the output as well.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.