Solving Constraint Integer Programs

bandit_epsgreedy.h File Reference

Detailed Description

internal methods for epsilon greedy bandit selection

Gregor Hendel

Definition in file bandit_epsgreedy.h.

#include "blockmemshell/memory.h"
#include "scip/def.h"
#include "scip/type_bandit.h"
#include "scip/type_retcode.h"
#include "scip/type_scip.h"

Go to the source code of this file.


SCIP_RETCODE SCIPincludeBanditvtableEpsgreedy (SCIP *scip)
SCIP_RETCODE SCIPbanditCreateEpsgreedy (BMS_BLKMEM *blkmem, BMS_BUFMEM *bufmem, SCIP_BANDITVTABLE *vtable, SCIP_BANDIT **epsgreedy, SCIP_Real *priorities, SCIP_Real eps, SCIP_Bool preferrecent, SCIP_Real decayfactor, int avglim, int nactions, unsigned int initseed)

Function Documentation

◆ SCIPincludeBanditvtableEpsgreedy()

SCIP_RETCODE SCIPincludeBanditvtableEpsgreedy ( SCIP scip)

creates the epsilon greedy bandit algorithm includes it in SCIP

scipSCIP data structure

Definition at line 339 of file bandit_epsgreedy.c.

References BANDIT_NAME, SCIP_CALL, SCIP_OKAY, and SCIPincludeBanditvtable().

Referenced by SCIPincludeCorePlugins().


SCIP_DECL_BANDITFREE ( SCIPbanditFreeEpsgreedy  )

callback to free bandit specific data structures

Definition at line 69 of file bandit_epsgreedy.c.

References BMSfreeBlockMemory, BMSfreeBlockMemoryArray, NULL, SCIP_OKAY, SCIPbanditGetData(), SCIPbanditGetNActions(), and SCIPbanditSetData().


SCIP_DECL_BANDITSELECT ( SCIPbanditSelectEpsgreedy  )


SCIP_DECL_BANDITUPDATE ( SCIPbanditUpdateEpsgreedy  )

update callback for bandit algorithm

Definition at line 155 of file bandit_epsgreedy.c.

References NULL, SCIP_OKAY, SCIP_Real, and SCIPbanditGetData().


SCIP_DECL_BANDITRESET ( SCIPbanditResetEpsgreedy  )

◆ SCIPbanditCreateEpsgreedy()

SCIP_RETCODE SCIPbanditCreateEpsgreedy ( BMS_BLKMEM blkmem,
BMS_BUFMEM bufmem,
SCIP_BANDIT **  epsgreedy,
SCIP_Real priorities,
SCIP_Real  eps,
SCIP_Bool  preferrecent,
SCIP_Real  decayfactor,
int  avglim,
int  nactions,
unsigned int  initseed 

internal method to create and reset epsilon greedy bandit algorithm

blkmemblock memory
bufmembuffer memory
vtablevirtual function table with epsilon greedy callbacks
epsgreedypointer to store the epsilon greedy bandit algorithm
prioritiesnonnegative priorities for each action, or NULL if not needed
epsparameter to increase probability for exploration between all actions
preferrecentshould the weights be updated in an exponentially decaying way?
decayfactorthe factor to reduce the weight of older observations if exponential decay is enabled
avglimnonnegative limit on observation number before the exponential decay starts, only relevant if exponential decay is enabled
nactionsthe positive number of possible actions
initseedinitial random seed

Definition at line 243 of file bandit_epsgreedy.c.

References BMSallocBlockMemory, BMSallocBlockMemoryArray, eps, NULL, SCIP_ALLOC, SCIP_CALL, SCIP_OKAY, and SCIPbanditCreate().

Referenced by SCIPcreateBanditEpsgreedy().