Single machine scheduling problem - Total Early Work

The benchmark set of instances is generated following instructions provided in the paper of Ben-Yehoshua and Mosheiov [1].

follows. Each instance is characterized by the number n of jobs to schedule, the maximum processing time of a job Pmax, and the density of due-dates alpha.
In the benchmark data set n takes value from the set {20, 50, 100, 200, 500, 1000, 10000}, Pmax varies in the set {50,100, 250, 500,1000}, while the tightness factor alpha takes value 0.7 or 1.0.
For each combination of n, Pmax, alpha values, 10 different instances are generated.
In each of ten instances, processing times of jobs are chosen uniformly from the interval [1, Pmax], while the due dates are chosen uniformly from the interval [1, alpha *SumP], where SumP is the sum of the processing times of the jobs.
The generated benchmark set is divided into three subsets according to the problem size/difficulty:
i) small instances with n <= 200,
ii) medium instances with n in {500, 1000} and
iii) large instances with n=10000.

REFERENCE :
[1] Ben-Yehoshua, Y., & Mosheiov, G. (2016). A single machine scheduling problem to minimize total early work. Computers & Operations Research, 73, 115-118.