A genetic algorithm (GA)-based model to deal with time-cost trade-off
problems is presented. The traditional algorithms assume the unlimited
availability of resources. Instead, the proposed model allows for res
ource constraints. Accordingly, the trade-off is considered in terms o
f the level of resources to be deployed for each activity. At the same
time the model schedules the starts of the activities in order to opt
imize the objective function. The activity starts is a significant fac
tor in the case of non-uniform available profile. The GA searches both
spaces of resource utilization and activity starts to determine the o
ptimal schedule that conforms to the resource available profile. Altho
ugh, in principle, this has the potential of a combinatorial explosion
. earlier work suggests that the GA-based model can be applied to larg
er networks without appearing to suffer from this problem. Allowing th
e project manager to think in terms of resource utilization makes this
approach consistent with resource allocation problems, and more pragm
atic and appealing. The model can also be used to solve conventional t
ime-cost trade-off problems by a simple modification of the objective
function. Numerical examples are used to illustrate the working of the
model. The examples include a case of varying resource available prof
ile.