The exputil test problem suite consists of 9 instances of a single-stage nonlinear stochastic 0-1 problem, specifically an expected utility knapsack problem. The problem formulation is described in [1].
The data for the instances are contained in the zip folder exputil.zip. Please refer to README.txt in the folder for problem formulation and data format. The instances and their optimal values are listed in the following table. The instances are named as exputil_n_m where n is the number of binary variables, and m is the number of scenarios.
|
Instance |
Optimal Value |
|
exputil_25_100 |
-0.242304 |
|
exputil_25_500 |
-0.246211 |
|
exputil_25_1000 |
-0.242750 |
|
exputil_50_100 |
-0.246343 |
|
exputil_50_500 |
-0.247751 |
|
exputil_50_1000 |
-0.247643 |
|
exputil_100_100 |
-0.247542 |
|
exputil_100_500 |
-0.248989 |
|
exputil_100_1000 |
-0.249317 |
REFERENCE:
[1] Shabbir Ahmed and Alper Atamturk. Maximizing a class of submodular utility functions. Mathematical Programming, 128(1-2):149–169, 2011.