July 2019
Consolidation of results:
- investigate why optimized rule-based control gives a rather superior performance than SDP (from which is was inspired): control discretization step too big?
New:
- create a Stochastic (tree-based) MPC Julia notebook
- evaluate performance uncertainty with bootstrapped inputs
- update the Comparison Python notebook with Julia MPC and OLFC
- update the Comparison Python notebook to latest
benchutils
functions
New sizing criterion: minimize grid energy cost + subscription cost? (because reducing subcription cost is one justification the use a battery)