[Previous] [Next]

maximize, minimize

Purpose

To optimize the stochastic problem.

Synopsis

maximize(obj: splinctr)

maximize(obj: splinctr, option: string)

maximize(obj: splinctr, t: integer, n: integer)

maximize(obj: splinctr, s: integer)

minimize(obj: splinctr)

minimize(obj: splinctr, option: string)

minimize(obj: splinctr, t: integer, n: integer)

minimize(obj: splinctr, s: integer)

Arguments

obj

Objective function

option  

Options for choosing combination of algorithms and types of problems to be solved

XSP_REC	Solve Recouse problem
XSP_EV	Solve Exprected Value problem
XSP_PI	Solve Perfect Information problem
XSP_OPT_LP_ONLY	Solve LP relaxation of the problem
XSP_PRIMAL	Solve problem using Primal algorithm
XSP_DUAL	Solve problem using Dual algorithm
XSP_BARRIER	Solve problem using Barrier algorithm

t

stage number belonging to {1..T}, where T is number of stages

n

node number belonging to {1..N_t}, where N_t is number of nodes in stage t

s

scenario number belonging to {1..S}, where S is number of scenarios

Further information

1. This procedure should be called after building scenario tree and stochastic constraints.

2. By default Recourse problem is solved.

3. Expected Value with Recourse can be solved using the option XSP_EV+XSP_REC.

4. Perfect Information with Recourse can be solved using the option XSP_PI+XSP_REC.

5. A problem instance in a particular scenario may be solved by passing the scenario number. The problem should be scenario based, i.e., xsp_scen_based should be true. User may want to prevent the overhead in this case by setting xsp_re_init to false.

6. A problem instance at a particular node in a scenario tree may be solved by passing the stage and the node number. The problem should be node based, i.e.. xsp_scen_based should be false. User may want to prevent the overhead in this case by setting xsp_re_init to false.

7. Hidden constraints are not considered during optimization. If there are some terms in any constraint containing fixed variables, they are substracted off from the constraint's right hand side.

8. Users should check the xprs_lpstatus, or xprs_mipstatus and compare it to constants XSP_LP_OPTIMAL, or XSP_MIP_OPTIMAL resp to ensure that the current problem is optimal.