Search space reduction in dynamic programming using monotonic heuristics in the context of model predictive optimization

Energy efficiency has become a major issue in trade, transportation and environment protection. While the next generation of zero emission propulsion systems are still under development, it is already possible to increase fuel efficiency in regular vehicles by applying a more fuel efficient driving behaviour. This paper proposes a model predictive A* optimization that makes use of a power-train model and the topography for the road ahead. The main scientific contribution is the development of admissible and monotonic non-trivial heuristics that allow A* to be used in an efficient manner while preserving global optimality. Simulations show that the heuristics guided optimization traverses a significantly smaller search space than dynamic programming without heuristics while preserving global optimality.