Abstract
We study the problem to simultaneously decide on the structures and the schedules for an entire portfolio of flexible projects. The projects are flexible as alternative technologies and procedures can be used to achieve the respective project task. The choice between different technologies and procedures affects the activities to be implemented and thus the precedence relations, i.e., the structure of the project. The different projects have given due dates with specific delay payments and compete for scarce resources. In this situation, project structure decisions and scheduling decisions are highly intertwined and have to be made simultaneously in order to achieve the assumed objective of minimizing the delay payments for the entire project portfolio. The problem is formally stated and solved via novel and problem-specific genetic algorithms. The performance of the new algorithms is evaluated with respect to speed and accuracy in a systematic and comprehensive numerical study.
Highlights
In this paper, we study the problem to structure and schedule multiple projects
As an alternative to using an activity list, we introduce a vector = ( l,j) of random keys which could look as follows in our example: Activity X-2 with a priority value of 0.7 is preferred to activity Y-3 with a value of only 0.1 and could be given priority when decoding the solution via the serial schedule generation scheme (SSGS) or the parallel schedule generation scheme (PSGS)
If we operate with the self-adapting schedule generation scheme (SGS) or the flexible serial-parallel SGS, we randomly determine whether the SSGS or the PSGS should be applied for the whole schedule or for each potentially possible scheduling step, respectively
Summary
We study the problem to structure and schedule multiple projects. The different projects have to be considered simultaneously as they compete for scarce resources such as human workers or technical equipment with exogenously limited capacities. The set of activities is not assumed to be exogenously given as alternative activities or entire sets of activities can serve the same purpose In such a situation, the project structure decision determines the activities that are implemented. Since genetic algorithms have frequently been proposed to determine highquality solutions for project scheduling problems quickly, we have developed genetic algorithms to make the two intertwined decisions on the project structures and schedules. This requires new and problem-specific representations and decoding procedures like those presented in this paper.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
