five

黑盒优化Benchmark——RABBO

收藏
阿里云天池2025-01-04 更新2024-03-07 收录
下载链接:
https://tianchi.aliyun.com/dataset/111953
下载链接
链接失效反馈
官方服务:
资源简介:
RABBO(Real-Application Black-Box Optimization benchmark)榜单提供具有实际应用背景的黑盒优化测试问题及评测方案,旨在帮助算法研发者打磨求解真实场景问题的黑盒优化算法,为算法使用者提供各类算法特点与适用场景分析以及使用参考。<br /> ● 黑盒优化问题,泛指目标函数难以从数学上解析表达,缺少可直接利用的梯度信息,仅可利用目标函数输入和对应输出函数值进行最优解搜索的优化问题<br /> ● RABBO提供了针对黑盒优化问题的数学建模和优化求解的思路,提供了黑盒优化接口规范代码、实际应用背景的测试问题、和效果评测的方案。帮助广大研发者快速学习和研发。<br /> ● RABBO榜单由达摩院决策智能实验室倾力维护,依托于阿里云天池平台的支持,将实行长期线上评测和定期揭榜。<br /> ● RABBO将持续拓展问题场景,为领域内研发团队提供一个“研发工具”和“竞技平台”,打造国内外最具活力的的黑盒优化Benchmark榜单。

The RABBO (Real-Application Black-Box Optimization Benchmark) provides black-box optimization test problems with practical application backgrounds and supporting evaluation protocols. It aims to assist algorithm developers in refining black-box optimization algorithms for solving real-world problems, and offer algorithm users analyses of algorithm characteristics, applicable scenarios, and practical usage references. - Black-box optimization problems refer to optimization tasks where the objective function has no tractable analytical mathematical expression, lacks directly accessible gradient information, and the optimal solution can only be searched by leveraging the input arguments of the objective function and their corresponding evaluated function values. - RABBO provides frameworks for mathematical modeling and optimal solution of black-box optimization problems, along with standardized interface code for black-box optimization, test problems with practical application backgrounds, and performance evaluation protocols, to help a broad range of researchers quickly learn and conduct their research and development work. - Maintained dedicatedly by the Decision Intelligence Lab of DAMO Academy with support from the Alibaba Cloud Tianchi Platform, RABBO will carry out long-term online evaluations and periodic leaderboard updates. - RABBO will continuously expand its problem scenarios, providing research and development teams in the field with a "research tool" and a "competition platform", and strive to build the most dynamic black-box optimization benchmark leaderboard both domestically and internationally.
提供机构:
阿里云天池
创建时间:
2021-10-14
搜集汇总
数据集介绍
main_image_url
以上内容由遇见数据集搜集并总结生成
二维码
社区交流群
二维码
科研交流群
商业服务