A hybrid heuristic–reinforcement learning method for multi-table join optimization in cloud-native databases
Multi-table join optimization in cloud-native databases faces challenges of storage-compute separation, resource elasticity, and runtime cost variation that classic optimizers cannot handle, which drives the need for methods able to adjust their decision boundaries according to the runtime behavior of the underlying cl...