Adjusted recombination operator for simulation-based construction schedule optimization

An efficient execution of complex construction projects requires a comprehensive scheduling of all construction activities. For this, it is necessary to consider various restrictions and the availability of required resources. The generation of efficient schedules is a very challenging task, in fact an NP-hard optimization problem. An appropriate approach is the application of discrete-event simulation for the generation of valid schedules. Additionally, simulation is combined with optimization methods to determine efficient schedules regarding given objectives. The applied optimization techniques are evolutionary algorithms. Thus, operators have to be implemented that define the way of generating new schedules in the recombination step of the algorithm. In this paper an improved operator is presented that outperforms common operators for scheduling problems by considering technological dependencies between activities, so that it will be possible to determine efficient schedules for complex construction scheduling problems. An example of implementation is presented to validate the developed operator.