This predicate finds start times for a set of tasks, which minimise the value of a given cost function.
The cost function is the only information the search algorithm can use to focus on the optimum. It cannot guide the search if the cost is a variable, only linked to the tasks start times by constraints. For this reason the cost function admits the special functions abs and maxlist.
The syntax for cost functions is:
CostFunction ::- PosExpr | PosExpr + PosExpr | Integer * PosExpr PosExpr ::- abs(LinearExpr) | maxlist([LinearExpr]) | LinearExpr.
The algorithm is described in more detail in the documentation of probe_cstr_sched/7.
probe_schedule(Starts,CostFun) :-
Starts=[X,Y,Z],
fd:(Starts::1..10),
Durations=[10,5,5],
Resources=[R1,R2,R3],
fd:(R1::1..2), R2=2, R3=1,
MaxResource=2,
[OldX,OldY,OldZ]=[1,5,5],
CostFun= abs(X-OldX)+abs(Y-OldY)+abs(Z-OldZ),
probe_sched(Starts,Durations,Resources,MaxResource,CostFun).