HDSDP is a numerical software solving the semidefinite programming problems. The main framework of HDSDP resembles the dual-scaling interior point solver DSDP[2] and several new features, especially a dual method based on the simplified homogeneous self-dual embedding, have been implemented. The embedding enhances stability of dual method and several new heuristics and computational techniques are designed to accelerate its convergence. HDSDP aims to show how dual-scaling algorithms benefit from the self-dual embedding and it is developed in parallel to DSDP5.8. Numerical experiments over several classical benchmark datasets exhibit its robustness and efficiency, and particularly its advantages on SDP instances featuring low-rank structure and sparsity. The pre-built binary of HDSDP is currently freely available at https://github.com/COPT-Public/HDSDP.
翻译:HDSDP是解决半无限期编程问题的一种数字软件,HDSDP的主要框架类似于双重缩放的内部点解解解码 DSDP[2],并且已经实施若干新的特点,特别是基于简化的单一自我嵌入的双重方法;嵌入可以增强双重方法的稳定性,并设计了几种新的休眠和计算技术以加速其趋同;HDSDP的目的是显示双重缩放算法如何从自我嵌入中受益,并与DSDP5.8.平行开发;对若干古典基准数据集的数值实验显示了其稳健和效率,特别是其对于以低级结构和宽度为特征的SDP实例的优势。HDSDP的预建二进制二进制目前可在https://github.com/COPT-Pub/HDSDP上自由查阅。