Test Instances. Cloud Software Group may offer test instances or environments, trials, or demos (collectively, “Test Instances”) as part of the Developer Materials, and you may use these Test Instances solely to develop and test your Application, subject to the other terms of this Agreement. These Test Instances are designed for short-term, temporary use and may be reset at any time without notice. You should not upload data or content to the Test Instances that is sensitive or confidential, or that you do not have intellectual property rights to use and upload.
Test Instances. There are 9 test instances that were created to test GT-EDD algorithm. Parameters for the test instances are shown in Table 2. Test instance, # Number of products Probability of a defective operation Demand, range, units Setup time, range, min Processing time, range, min Due date, range, days Buffer capacity, units Buffer time, min 1 30 30% [10, 20] [10, 20] [2, 10] [1, 3] 2 20 2 30 30% [10, 20] [10, 20] [2, 10] [1, 3] 3 20 3 30 30% [10, 20] [10, 20] [2, 10] [1, 3] 5 20 4 40 30% [10, 20] [10, 20] [2, 10] [1, 3] 2 20 5 40 30% [10, 20] [10, 20] [2, 10] [1, 3] 3 20 6 40 30% [10, 20] [10, 20] [2, 10] [1, 3] 5 20 7 50 30% [10, 20] [10, 20] [2, 10] [1, 3] 2 20 8 50 30% [10, 20] [10, 20] [2, 10] [1, 3] 3 20 9 50 30% [10, 20] [10, 20] [2, 10] [1, 2] 5 20 Test instances are divided into 3 groups of 30, 40 and 50 products. Each group consists of 3 instances. These instances differ from each other only by the buffer capacity which takes the values of 2, 3, or 5 units. Probability of a defective operation and a fixed time that any item spends in the buffer denoted in Table 2 as buffer time are held constant for all instances. Demand, setup times, processing times and due dates are randomly generated within the ranges specified in Table 2. Due dates vary between 1 and 3 days in minutes. Notice that buffer time is twice larger than the largest processing time. Moreover, the generated G(g,f,r) function has not a concave staircase structure in
Test Instances. Test instances used in this paper consists of two sets of single item instances generated by Xxxxxxxx et al. (2005). Both sets contains 96 instances divided into four classes. The first set has T = 6 time periods, and N = 50,75,100,125,150 customers. The second set have T = 30 time periods and N = 50 customers. The different classes have different costs for the given parameters to emulate different situations in real world examples: unitary production cost cp, fixed production cost per period fp, inventory holding cost at customer k hk and fixed transportation cost per period fv. In class 1 these parameters are set to h0 = 3, cp = 10h0, fp = 100cp and fv = 32181. For evaluation of situations where there are no fixed transportation costs specified, class 2 has fixed transportation cost fv = 0. To see what impact on the solution there will be if the production cost is reduced compared to the inventory holding cost at the plant, the instances in class 3 have a unit production cost cp = h0. In the last class the customer inventory holding cost hk, k = 1, ..., N are set to zero. This will emulate situations where the plant does not have to pay for customer inventory holding costs. These instances have been applied by both Xxxxxxxx and Xxxxxxxxxx in their papers about the IPDP. Xxxxxxxx used the more constrained VMI-OU and VMI-FFD strategies in his paper, while Xxxxxxxxxx used the less constrained VMI-ML strategy to measure the cost reduction this strategy gives over the results reported by Xxxxxxxx. As mentioned earlier the tests performed in this thesis uses the VMI-ML strategy which give a god basis for result comparizon against results reported by Xxxxxxxxxx. Intel Core 2 Duo , 3 Ghz Intel Quad i7 920, 3.66 Ghz MacBook Pro 2.2 Ghz Core 2 Duo
