Solvers Chaining in the IDOS Server for Dynamic Optimization

Pytlak, R; Błaszczyk, J; Karbowski, A; Krawczyk, K; Tarnawski, T

  • the 52nd IEEE Conference on Decision and Control (CDC 2013);
  • Tom: -;
  • Strony: 7119-7124;
  • 2013;

This paper presents a newly deployed server, IDOS (Interactive Dynamic Optimization Sever), devoted to solving optimal control problems. Development and deployment of the Interactive Dynamic Optimization Server is a result of a project funded by NCBiR (National Center for Research and Development in Poland). The aim of the project was to develop a prototype, online-accessible environment for solving dynamic optimization problems.Within the project we also constructed a modeling language (Dynamic Optimization Modeling Language, DOML) for defining optimal control problems. In result, a user can describe his problem in a programming-language independent way. Then, once defined, the problem can be attempted by different solvers. In particular one can get a crude approximation to a solution of his problem by applying one solver and then continue solving the problem by another solver which gives a more accurate solution but may require a better initial guess than the first solver. The paper describes some constructs of DOML language which enable such chaining of solvers. A simple optimal control problem is used to illustrate the functionality of the IDOS server and the completeness of of our language DOML.