On December 28, it was reported that at the first energy electronics industry innovation competition sponsored by the Industrial Development Promotion Center of the Ministry of Industry and Information Technology and other units, the self-developed optimization solver mindopt of Ali Damo Academy (Lakeside Laboratory) won the first place in the domestic solver for electric power. The solver, also known as the "core of industrial software", is related to the national economy and people's livelihood industries such as energy and power, transportation and logistics. After four years and 26 iterations, the Damo Academy has created a self-developed solver with complete functions and world-class performance.
As the underlying industrial software, the optimization solver is widely used in fields that require complex calculation and planning, such as energy and power, industrial manufacturing, transportation and logistics, with high technical barriers and difficult research and development, and has been dominated by European and American companies for decades. Among them, power involves the national economy and people's livelihood and strategic security, and the optimization solution of power grid dispatching operation has the characteristics of many variables, complex constraints and difficult solutions, which urgently needs the development and breakthrough of domestic independent controllable solvers. Therefore, this competition has set up a domestic solver problem, the content is based on the power system unit combination and economic scheduling problems as the background, the decision variables are the unit start-up and shutdown and the unit output, and the constraints include unit operation constraints, system balance constraints, power grid safety constraints, etc., so as to minimize the start-up and shutdown costs and operating costs of the units.
According to the organizers, this competition brings together all the domestic commercial optimization solver development teams and some power dispatching research institutes. After two rounds of test matches and one round of official competitions, the Ali Damo Academy solver finally won the first prize. In addition, Tsinghua University and Qingneng Internet United, which won the second prize of this competition, also provided some technical support by the self-developed solver of the DAMO Academy. According to reports, the Damo Academy solver has excellent performance in accuracy index and speed index, which lies in the comprehensive use of innovative technologies such as heuristic neighborhood search algorithm based on relaxation degenerate movement, linear constraint propagation with maximum group enhancement, and successfully grasping and utilizing the special structure of electric power. Last year, the DAMO Academy's self-developed solver team also won the championship in international power and energy competitions such as the Neurips Virtual Power Plant Competition and GECCO&IEEE Energy Dispatching.
It is understood that under the leadership of the internationally renowned mathematician Yin Wotao, the Decision Intelligence Laboratory of Ali Damo Academy has been committed to the research and development of solvers since 2019, starting from scratch, after four years and 26 iterations, and finally officially launched V1 in October 2023Version 0, becoming a self-developed solver with complete functions and leading performance. At present, the DAMO Academy solver has been able to solve mainstream optimization problems including linear programming, mixed integer linear programming, large-scale network flow, convex quadratic programming, semidefinite programming, and nonlinear programming, and has the ability to accelerate AI for focused fields.
At the same time, the DAMO Academy has also continuously broken through the original framework of the solver, combined with the actual needs of China's industrial scenarios, and successively launched its own modeling language, parameter tuner, and optimization platform. In August this year, we built a large model for decision-making and reasoning, and launched Mindopt Copilot, an AI engineer with mathematical optimization technical capabilities. Users can directly use natural language to ask optimization questions without complex mathematical or programming knowledge, and Mindopt Copilot can automatically mathematical modeling, transform the problem into optimization models for linear programming and mixed integer linear programming, and code and call the software to calculate the best answer. Yin Wotao, head of the Damo Academy's Decision Intelligence Laboratory, said: "Since its start four years ago, the Damo Academy solver has set its goal as the next generation of technology, rather than just catching up with international leading manufacturers. ”
At present, the self-developed solver of Ali Damo Academy has been widely used in cloud computing, ** chain, e-commerce, finance and other fields. Especially in the field of power and energy, the DAMO Academy has cooperated with the State Grid of China to reach the international leading level in performance on a number of power dispatching issuesIn cooperation with the Power Dispatching and Control Center of China Southern Power Grid, the "Power Dispatching Intelligent Decision-making Platform" was released to help the Southern Power Grid General Dispatcher achieve dispatching from 15 minutes to seconds, with an accuracy rate of experienced dispatchers, and supported the 4th and 5th China Southern Power Grid Power Dispatching AI Application Competition.
Disclaimer: The content of this article is this **publication or **corporate promotional information, which only represents the author's personal views and has nothing to do with this website. It is for readers' reference only, and please verify the relevant content on your own.