Using computer algebra for solving some optimal control problems by dynamic programming

This paper deals with the application of dynamic programming to control problems by using computer algebra. Computer algebra has been principally used here for solving dynamic programming functional equations. This approach is particularly suitable for non-linear problems. Practical implementation on a PC has been done by using MapleV software; some numerical examples are given. It has been found that for small problems the method worked well even for highly non-linear problems. However, for large-scale non-linear ones the hardware and software combination used was unable to reach solutions. Further investigations would be carried out to push back the limiting factor.