Track: Information Technology and Information Systems

Abstract

Gupta algorithm was proposed back in 1972 to reduce cycle times, with very good performance results and it has been successfully used as a sequencing solution in critical activities in lean software develop processes. An alternative to this algorithm is presented with validation evidence on this document, using heuristic rules when equal selecting factors occurs. The algorithm, named Caldwell-T, can be used to solve, effectively, the possible clashes of activity that can occur in shared resources, as opposed to the traditional methodology, which encourages multitasking. Caldwell-t Algorithm obtained 78% performance effectiveness over Gupta proposal using a simulation model with Industry data. This improvement gives a big impact in software develop projects, when it is applied to critical chains and feeding buffers and signal calculations.

However, the new algorithm may yield less effective results compared with Gupta algorithm for certain random configurations of expected times of execution of activities, in approximately 20% of the cases, which allows to deduce that it is better to analyze the configuration of the multiproject environment with both algorithms, instead of putting them to compete. The results of this research is highly valuable because Gupta Algorithm is powerful compare with GA solutions and other IA approaches.