java.lang.Object
org.cicirello.search.problems.scheduling.WeightedShortestProcessingPlusSetupTime
All Implemented Interfaces:
Splittable<ConstructiveHeuristic<Permutation>>, ConstructiveHeuristic<Permutation>

public final class WeightedCriticalRatioSetupAdjusted
This is an implementation of a variation of the weighted critical ratio heuristic, adjusted to account for setup times for problems with sequence-dependent setups. The usual definition of this heuristic is: h(j) = (w[j]/p[j])(1/(1+S(j)/p[j])), where w[j] is the weight of job j, p[j] is its processing time, and S(j) is a calculation of the slack of job j where slack S(j) is d[j] - T - p[j] - s[i][j]. The d[j] is the job's due date, T is the current time, and s[i][j] is setup time of the job if it follows job i (for problems with setup times).

Historically, this heuristic has been criticized for allowing negative evaluations (i.e., slack S(j) is negative for jobs completing late). Additionally, this library's use of constructive heuristics is for stochastic sampling, for which we require positive heuristic values. Therefore, we have altered the definition as follows: h(j) = (w[j]/p[j])(1/(1+max(0,S(j))/p[j])).

Finally, to adjust for setup times, we replace p[j] wherever it appears with p[j]+s[i][j]. This leads to: h(j) = (w[j]/(p[j]+s[i][j]))(1/(1+max(0,S(j))/(p[j]+s[i][j]))).

Furthermore, the constant MIN_H defines the minimum value the heuristic will return, preventing h(j)=0 in support of stochastic sampling algorithms for which h(j)=0 is problematic. This implementation returns max( MIN_H, h(j)), where MIN_H is a small non-zero value.

• Field Details

• MIN_H

public static final double MIN_H
The minimum heuristic value. If the heuristic value as calculated is lower than MIN_H, then MIN_H is used as the heuristic value. The reason is related to the primary purpose of the constructive heuristics in the library: heuristic guidance for stochastic sampling algorithms, which assume positive heuristic values (e.g., an h of 0 would be problematic).