Cubes design in maintenance data warehouse with genetic algorithms

Nowadays, maintenance management has been one of the most important tasks in equipment management, particularly in manufacturing industries. Equipment Management System (EMS) aims at reducing maintenance cost and production loss caused by machine breakdown. As well, EMS can assist equipment engineers to make the right maintenance decisions at the right time, and at the right shop floor. In this paper, we design a data warehouse for EMS to help equipment engineers make maintenance decisions with various equipment related dimens ions in an effective manner. Several cubes such as Mean Time To Repair (MTTR), Mean Time Between Failure (MTBF), Spare Part Response Time (SPRT), etc. can be built from EMS data warehouse for the purpose of decisionmaking. In order to achieve a reasonable query response time under the memory space limit , partial materialization is adopted to cube selection. This paper compares the optimization approach based on Genetic Algorithms (GAs) to the traditional greedy search for designing data cubes in EMS data warehouse. According to the experimental results, the partial materialization suggested by the GAbased approach can decrease 70% of memory space of full materialization, while the query response times are approximately unchanged.