This repository contains AMPL implementations of the integer linear program specified in the paper Self-Driving Database Systems: A Conceptual Approach. In their current state, the programs depend on AMPL and the Gurobi solver. After starting AMPL, run the programs with include PROGRAM_NAME.txt
This program contains code that determines an efficient tuning order for index selection, compression selection and clustering (Table 1 and 2 from the paper). Workload costs are determined experimentally for the main memory research database Hyrise.
The second program contains the code to generate data for the scalability examples (Example 1), and thereby the results for Table 3, 4, and 5.