Download LINDO API USER MANUAL

Transcript
STOCHASTIC PROGRAMMING
D
R
0
1
TIME2
TIME2
405
1
1
After the new time structure is loaded to LINDO API, we can work out the loading of the stochastic
data as follows. Suppose the joint distribution probabilities are the same as case 3. This leads to the
following scenario tree.
Note:
This new version of Newsvendor problem is actually a special case of the original problem,
where stage 1 and stage 2 (TIME2 and TIME3) are aggregated into a single stage which is
now called stage 1 (TIME2). The consequences of this aggregation are that all random
parameters, constraints and variables that belonged to stage 1 and stage 2 in the original
problem now belong to stage 1 in the aggregated version.
As it can be seen in the scenario tree, each outcome in stage 1 corresponds to a block realization of a
vector of random parameters, namely D and R. The associated stochastic data can be loaded to LINDO
API as in the following code snippet in C language. See lindoapi/samples/c/ex_sp_newsboy directory
for the complete application modeling this case.