weightedAStar.cpp

Go to the documentation of this file.

/*
 * Copyright (c) 2008, Maxim Likhachev
 * All rights reserved.
 * 
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are met:
 * 
 *     * Redistributions of source code must retain the above copyright
 *       notice, this list of conditions and the following disclaimer.
 *     * Redistributions in binary form must reproduce the above copyright
 *       notice, this list of conditions and the following disclaimer in the
 *       documentation and/or other materials provided with the distribution.
 *     * Neither the name of the University of Pennsylvania nor the names of its
 *       contributors may be used to endorse or promote products derived from
 *       this software without specific prior written permission.
 * 
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
 * POSSIBILITY OF SUCH DAMAGE.
 */
#include <iostream>
using namespace std;

#include <sbpl/headers.h>
#include <sbpl_dynamic_planner/weightedAStar.h>



//-----------------------------------------------------------------------------------------------------

WeightedAStar::WeightedAStar(DiscreteSpaceTimeInformation* environment, bool bSearchForward)
{
  WeightedAStar_FINAL_EPS = 1.0;
        bforwardsearch = bSearchForward;

    environment_ = environment;
    
        bsearchuntilfirstsolution = false;
    finitial_eps = WeightedAStar_DEFAULT_INITIAL_EPS;
    searchexpands = 0;
    MaxMemoryCounter = 0;
    
    fDeb = SBPL_FOPEN("debug.txt", "w");
    
    pSearchStateSpace_ = new WeightedAStarSearchStateSpace_t;
    
    
    //create the WeightedAStar planner
    if(CreateSearchStateSpace(pSearchStateSpace_) != 1)
        {
            SBPL_ERROR("ERROR: failed to create statespace\n");
            return;
        }
    
    //set the start and goal states
    if(InitializeSearchStateSpace(pSearchStateSpace_) != 1)
        {
            SBPL_ERROR("ERROR: failed to create statespace\n");
            return;
        }    
}

WeightedAStar::~WeightedAStar()
{
  if(pSearchStateSpace_ != NULL){
    //delete the statespace
    DeleteSearchStateSpace(pSearchStateSpace_);
    delete pSearchStateSpace_;
  }
  SBPL_FCLOSE(fDeb);
}


void WeightedAStar::Initialize_searchinfo(CMDPSTATE* state, WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{

        WeightedAStarState* searchstateinfo = (WeightedAStarState*)state->PlannerSpecificData;

        searchstateinfo->MDPstate = state;
        InitializeSearchStateInfo(searchstateinfo, pSearchStateSpace); 
}


CMDPSTATE* WeightedAStar::CreateState(int stateID, WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{       
        CMDPSTATE* state = NULL;

#if DEBUG
        if(environment_->StateID2IndexMapping[stateID][ARAMDP_STATEID2IND] != -1)
        {
                SBPL_ERROR("ERROR in CreateState: state already created\n");
                throw new SBPL_Exception();
        }
#endif

        //adds to the tail a state
        state = pSearchStateSpace->searchMDP.AddState(stateID);

        //remember the index of the state
        environment_->StateID2IndexMapping[stateID][ARAMDP_STATEID2IND] = pSearchStateSpace->searchMDP.StateArray.size()-1;

#if DEBUG
        if(state != pSearchStateSpace->searchMDP.StateArray[environment_->StateID2IndexMapping[stateID][ARAMDP_STATEID2IND]])
        {
                SBPL_ERROR("ERROR in CreateState: invalid state index\n");
                throw new SBPL_Exception();
        }
#endif


        //create search specific info
        state->PlannerSpecificData = (WeightedAStarState*)malloc(sizeof(WeightedAStarState));   
        Initialize_searchinfo(state, pSearchStateSpace);
        MaxMemoryCounter += sizeof(WeightedAStarState);

        return state;

}

CMDPSTATE* WeightedAStar::GetState(int stateID, WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{       

        if(stateID >= (int)environment_->StateID2IndexMapping.size())
        {
          SBPL_ERROR("ERROR int GetState: stateID %d is invalid\n", stateID);
                throw new SBPL_Exception();
        }

        if(environment_->StateID2IndexMapping[stateID][ARAMDP_STATEID2IND] == -1)
                return CreateState(stateID, pSearchStateSpace);
        else
                return pSearchStateSpace->searchMDP.StateArray[environment_->StateID2IndexMapping[stateID][ARAMDP_STATEID2IND]];

}



//-----------------------------------------------------------------------------------------------------




int WeightedAStar::ComputeHeuristic(CMDPSTATE* MDPstate, WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
        //compute heuristic for search

        if(bforwardsearch)
        {

#if MEM_CHECK == 1
                //int WasEn = DisableMemCheck();
#endif

                //forward search: heur = distance from state to searchgoal which is Goal WeightedAStarState
                int retv =  environment_->GetGoalHeuristic(MDPstate->StateID);

#if MEM_CHECK == 1
                //if (WasEn)
                //      EnableMemCheck();
#endif

                return retv;

        }
        else
        {
                //backward search: heur = distance from searchgoal to state
                return environment_->GetStartHeuristic(MDPstate->StateID);
        }
}


//initialization of a state
void WeightedAStar::InitializeSearchStateInfo(WeightedAStarState* state, WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
        state->g = INFINITECOST;
        state->v = INFINITECOST;
        state->iterationclosed = 0;
        state->callnumberaccessed = pSearchStateSpace->callnumber;
        state->bestnextstate = NULL;
        state->costtobestnextstate = INFINITECOST;
        state->heapindex = 0;
        state->listelem[WeightedAStar_INCONS_LIST_ID] = 0;
        state->numofexpands = 0;

        state->bestpredstate = NULL;

        //compute heuristics
#if USE_HEUR
        if(pSearchStateSpace->searchgoalstate != NULL)
                state->h = ComputeHeuristic(state->MDPstate, pSearchStateSpace); 
        else 
                state->h = 0;
#else
        state->h = 0;
#endif


}



//re-initialization of a state
void WeightedAStar::ReInitializeSearchStateInfo(WeightedAStarState* state, WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
        state->g = INFINITECOST;
        state->v = INFINITECOST;
        state->iterationclosed = 0;
        state->callnumberaccessed = pSearchStateSpace->callnumber;
        state->bestnextstate = NULL;
        state->costtobestnextstate = INFINITECOST;
        state->heapindex = 0;
        state->listelem[WeightedAStar_INCONS_LIST_ID] = 0;
        state->numofexpands = 0;

        state->bestpredstate = NULL;

        //compute heuristics
#if USE_HEUR

        if(pSearchStateSpace->searchgoalstate != NULL)
        {
                state->h = ComputeHeuristic(state->MDPstate, pSearchStateSpace); 
        }
        else 
                state->h = 0;

#else

        state->h = 0;

#endif


}



void WeightedAStar::DeleteSearchStateData(WeightedAStarState* state)
{
        //no memory was allocated
        MaxMemoryCounter = 0;
        return;
}



//used for backward search
void WeightedAStar::UpdatePreds(WeightedAStarState* state, WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
    vector<int> PredIDV;
    vector<int> CostV;
        CKey key;
        WeightedAStarState *predstate;

    environment_->GetPreds(state->MDPstate->StateID, &PredIDV, &CostV);

        //iterate through predecessors of s
        for(int pind = 0; pind < (int)PredIDV.size(); pind++)
        {
                CMDPSTATE* PredMDPState = GetState(PredIDV[pind], pSearchStateSpace);
                predstate = (WeightedAStarState*)(PredMDPState->PlannerSpecificData);
                if(predstate->callnumberaccessed != pSearchStateSpace->callnumber)
                        ReInitializeSearchStateInfo(predstate, pSearchStateSpace);

                //see if we can improve the value of predstate
                if(predstate->g > state->v + CostV[pind])
                {
                        predstate->g = state->v + CostV[pind];
                        predstate->bestnextstate = state->MDPstate;
                        predstate->costtobestnextstate = CostV[pind];

                        //re-insert into heap if not closed yet
                        if(predstate->iterationclosed != pSearchStateSpace->searchiteration)
                        {
                                key.key[0] = predstate->g + (int)(pSearchStateSpace->eps*predstate->h);
                                //key.key[1] = predstate->h;
                                if(predstate->heapindex != 0)
                                        pSearchStateSpace->heap->updateheap(predstate,key);
                                else
                                        pSearchStateSpace->heap->insertheap(predstate,key);
                        }
                        //take care of incons list
                        else if(predstate->listelem[WeightedAStar_INCONS_LIST_ID] == NULL)
                        {
                                pSearchStateSpace->inconslist->insert(predstate, WeightedAStar_INCONS_LIST_ID);
                        }
                }
        } //for predecessors

}

//used for forward search
void WeightedAStar::UpdateSuccs(WeightedAStarState* state, WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
    vector<int> SuccIDV;
    vector<int> CostV;
        CKey key;
        WeightedAStarState *succstate;

  vector<int> holdList;
  vector<int> activeBubbles;
  
  int tempX,tempY,tempTH,tempT;
  environment_->GetCoordFromState(state->MDPstate->StateID,tempX,tempY,tempTH,tempT);
  //if(state->v != tempT)
    //SBPL_PRINTF("expanded time (%d) not equal to cost (%d)\n",tempT,state->v);

  environment_->GetSuccs(state->MDPstate->StateID, &SuccIDV, &CostV, &holdList, &activeBubbles);

  //activate all bubbles that had collisions by inserting all of the on hold states into the open list
  for(unsigned int i=0; i<activeBubbles.size(); i++){
    //SBPL_PRINTF("Activate bubble %d\n", activeBubbles[i]);
    bubbles[activeBubbles[i]].active = true;
    //flush any states on the hold list
    for(unsigned int j=0; j<bubbles[activeBubbles[i]].hold.size(); j++){
      if(!bubbles[activeBubbles[i]].hold[j]->done){
        //SBPL_PRINTF("inserting %d\n",bubbles[activeBubbles[i]].hold[j]->state->MDPstate->StateID);
        //if(bubbles[activeBubbles[i]].hold[j]->state->MDPstate->StateID == 10602)
          //SBPL_PRINTF("being flushed...\n");
        pSearchStateSpace->heap->insertheap(bubbles[activeBubbles[i]].hold[j]->state, bubbles[activeBubbles[i]].hold[j]->key);
        //lazy deletion...
        bubbles[activeBubbles[i]].hold[j]->done = true;
      }
    }
  }

        //iterate through successors of s
        for(int sind = 0; sind < (int)SuccIDV.size(); sind++)
        {
                CMDPSTATE* SuccMDPState = GetState(SuccIDV[sind], pSearchStateSpace);
                int cost = CostV[sind];

                succstate = (WeightedAStarState*)(SuccMDPState->PlannerSpecificData);
                if(succstate->callnumberaccessed != pSearchStateSpace->callnumber)
                        ReInitializeSearchStateInfo(succstate, pSearchStateSpace);

                //see if we can improve the value of succstate
                //taking into account the cost of action
                if(succstate->g > state->v + cost)// && succstate->iterationclosed != pSearchStateSpace->searchiteration) //MIKE: added the second condition to prevent time from being different than cost
                {
                        succstate->g = state->v + cost;
                        succstate->bestpredstate = state->MDPstate; 

                        //re-insert into heap if not closed yet
                        if(succstate->iterationclosed != pSearchStateSpace->searchiteration)
                        {
                                //relax this hash entry's time
        environment_->Relax(state->MDPstate->StateID, SuccIDV[sind]);

                                key.key[0] = succstate->g + (int)(pSearchStateSpace->eps*succstate->h);

                                //key.key[1] = succstate->h;

                                if(succstate->heapindex != 0)
                                        pSearchStateSpace->heap->updateheap(succstate,key);
                                else{
          if(sind == (int)SuccIDV.size()-1 && holdList.size() > 0){
            bool onHold = true;
            for(unsigned int i=0; i<holdList.size(); i++)
              onHold = onHold && !bubbles[holdList[i]].active;
            if(onHold){
              //if(succstate->MDPstate->StateID == 10602)
                //SBPL_PRINTF("on hold we go...\n");
              //this wait state should be put on hold for later...
              //SBPL_PRINTF("On hold %d\n", SuccIDV[sind]);
              HoldState_t* temp = new HoldState_t;
              temp->state = succstate;
              temp->key = key;
              temp->done = false;
              for(unsigned int i=0; i<holdList.size(); i++)
                bubbles[holdList[i]].hold.push_back(temp);
            }
            else{
              //if(succstate->MDPstate->StateID == 10602)
                //SBPL_PRINTF("wait getting inserted...\n");
              pSearchStateSpace->heap->insertheap(succstate,key);
            }
          }
          else{
            //if(succstate->MDPstate->StateID == 10602)
              //SBPL_PRINTF("action getting inserted...\n");
            pSearchStateSpace->heap->insertheap(succstate,key);
          }
        }
                        }
                        //take care of incons list
                        else if(succstate->listelem[WeightedAStar_INCONS_LIST_ID] == NULL)
                        {
                                pSearchStateSpace->inconslist->insert(succstate, WeightedAStar_INCONS_LIST_ID);
                        }
                } //check for cost improvement 

        } //for actions
}

//TODO-debugmax - add obsthresh and other thresholds to other environments in 3dkin
int WeightedAStar::GetGVal(int StateID, WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
         CMDPSTATE* cmdp_state = GetState(StateID, pSearchStateSpace);
         WeightedAStarState* state = (WeightedAStarState*)cmdp_state->PlannerSpecificData;
         return state->g;
}

//returns 1 if the solution is found, 0 if the solution does not exist and 2 if it ran out of time
int WeightedAStar::ImprovePath(WeightedAStarSearchStateSpace_t* pSearchStateSpace, double MaxNumofSecs)
{
        int expands;
        WeightedAStarState *state, *searchgoalstate;
        CKey key, minkey;
        CKey goalkey;

        expands = 0;


        if(pSearchStateSpace->searchgoalstate == NULL)
        {
                SBPL_ERROR("ERROR searching: no goal state is set\n");
                throw new SBPL_Exception();
        }

        //goal state
        searchgoalstate = (WeightedAStarState*)(pSearchStateSpace->searchgoalstate->PlannerSpecificData);
        if(searchgoalstate->callnumberaccessed != pSearchStateSpace->callnumber)
                ReInitializeSearchStateInfo(searchgoalstate, pSearchStateSpace);

        //set goal key
        goalkey.key[0] = searchgoalstate->g;
        //goalkey.key[1] = searchgoalstate->h;

        //expand states until done
        minkey = pSearchStateSpace->heap->getminkeyheap();
        CKey oldkey = minkey;
        while(!pSearchStateSpace->heap->emptyheap() && minkey.key[0] < INFINITECOST && goalkey > minkey &&
                (clock()-TimeStarted) < MaxNumofSecs*(double)CLOCKS_PER_SEC) 
    {

                //get the state         
                state = (WeightedAStarState*)pSearchStateSpace->heap->deleteminheap();


#if DEBUG
                //SBPL_FPRINTF(fDeb, "expanding state(%d): h=%d g=%u key=%u v=%u iterc=%d callnuma=%d expands=%d (g(goal)=%u)\n",
                //      state->MDPstate->StateID, state->h, state->g, state->g+(int)(pSearchStateSpace->eps*state->h), state->v, 
                //      state->iterationclosed, state->callnumberaccessed, state->numofexpands, searchgoalstate->g);
                SBPL_FPRINTF(fDeb, "expanding: ");
                PrintSearchState(state, fDeb);
                if(state->listelem[WeightedAStar_INCONS_LIST_ID]  != NULL)
                {
                        SBPL_FPRINTF(fDeb, "ERROR: expanding a state from inconslist\n");
                        SBPL_ERROR("ERROR: expanding a state from inconslist\n");
                        throw new SBPL_Exception();
                }
                fflush(fDeb);
#endif

#if DEBUG
                if(minkey.key[0] < oldkey.key[0] && fabs(this->finitial_eps - 1.0) < ERR_EPS)
                {
                        //SBPL_PRINTF("WARN in search: the sequence of keys decreases\n");
                        //throw new SBPL_Exception();
                }
                oldkey = minkey;
#endif

                if(state->v == state->g)
                {
                        SBPL_ERROR("ERROR: consistent state is being expanded\n");
#if DEBUG
                        SBPL_FPRINTF(fDeb, "ERROR: consistent state is being expanded\n");
                        throw new SBPL_Exception();
#endif
                }

                //recompute state value      
                state->v = state->g;
                state->iterationclosed = pSearchStateSpace->searchiteration;

                //new expand      
                expands++;
                state->numofexpands++;


                if(bforwardsearch == false)
                        UpdatePreds(state, pSearchStateSpace);
                else
                        UpdateSuccs(state, pSearchStateSpace);
                
                //recompute minkey
                minkey = pSearchStateSpace->heap->getminkeyheap();

                //recompute goalkey if necessary
                if(goalkey.key[0] != (int)searchgoalstate->g)
                {
                        //SBPL_PRINTF("re-computing goal key\n");
                        //recompute the goal key (heuristics should be zero)
                        goalkey.key[0] = searchgoalstate->g;
                        //goalkey.key[1] = searchgoalstate->h;
                }

                if(expands%100000 == 0 && expands > 0)
                {
                        SBPL_PRINTF("expands so far=%u\n", expands);
                }

        }

        int retv = 1;
        if(searchgoalstate->g == INFINITECOST && pSearchStateSpace->heap->emptyheap())
        {
                SBPL_PRINTF("solution does not exist: search exited because heap is empty\n");
                retv = 0;
        }
        else if(!pSearchStateSpace->heap->emptyheap() && goalkey > minkey)
        {
                SBPL_PRINTF("search exited because it ran out of time\n");
                retv = 2;
        }
        else if(searchgoalstate->g == INFINITECOST && !pSearchStateSpace->heap->emptyheap())
        {
                SBPL_PRINTF("solution does not exist: search exited because all candidates for expansion have infinite heuristics\n");
                retv = 0;
        }
        else
        {
                SBPL_PRINTF("search exited with a solution for eps=%.3f\n", pSearchStateSpace->eps);
                retv = 1;
        }

        //SBPL_FPRINTF(fDeb, "expanded=%d\n", expands);

        searchexpands += expands;

        return retv;            
}


void WeightedAStar::BuildNewOPENList(WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
        WeightedAStarState *state;
        CKey key;
        CHeap* pheap = pSearchStateSpace->heap;
        CList* pinconslist = pSearchStateSpace->inconslist; 
                
        //move incons into open
        while(pinconslist->firstelement != NULL)
          {
            state = (WeightedAStarState*)pinconslist->firstelement->liststate;
            
            //compute f-value
            key.key[0] = state->g + (int)(pSearchStateSpace->eps*state->h);
            //key.key[1] = state->h;
            
            //insert into OPEN
            pheap->insertheap(state, key);
            //remove from INCONS
            pinconslist->remove(state, WeightedAStar_INCONS_LIST_ID);
          }
}


void WeightedAStar::Reevaluatefvals(WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
        CKey key;
        int i;
        CHeap* pheap = pSearchStateSpace->heap;
        
        //recompute priorities for states in OPEN and reorder it
        for (i = 1; i <= pheap->currentsize; ++i)
          {
                WeightedAStarState* state = (WeightedAStarState*)pheap->heap[i].heapstate;
            pheap->heap[i].key.key[0] = state->g + 
              (int)(pSearchStateSpace->eps*state->h); 
            //pheap->heap[i].key.key[1] = state->h; 
          }
        pheap->makeheap();

        pSearchStateSpace->bReevaluatefvals = false;
}




//creates (allocates memory) search state space
//does not initialize search statespace
int WeightedAStar::CreateSearchStateSpace(WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{

        //create a heap
        pSearchStateSpace->heap = new CHeap;
        pSearchStateSpace->inconslist = new CList;
        MaxMemoryCounter += sizeof(CHeap);
        MaxMemoryCounter += sizeof(CList);

        pSearchStateSpace->searchgoalstate = NULL;
        pSearchStateSpace->searchstartstate = NULL;

        searchexpands = 0;


    pSearchStateSpace->bReinitializeSearchStateSpace = false;
        
        return 1;
}

//deallocates memory used by SearchStateSpace
void WeightedAStar::DeleteSearchStateSpace(WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
        if(pSearchStateSpace->heap != NULL)
        {
                pSearchStateSpace->heap->makeemptyheap();
                delete pSearchStateSpace->heap;
                pSearchStateSpace->heap = NULL;
        }

        if(pSearchStateSpace->inconslist != NULL)
        {
                pSearchStateSpace->inconslist->makeemptylist(WeightedAStar_INCONS_LIST_ID);
                delete pSearchStateSpace->inconslist;
                pSearchStateSpace->inconslist = NULL;
        }

        //delete the states themselves
        int iend = (int)pSearchStateSpace->searchMDP.StateArray.size();
        for(int i=0; i < iend; i++)
        {
                CMDPSTATE* state = pSearchStateSpace->searchMDP.StateArray[i];
    if(state != NULL && state->PlannerSpecificData != NULL){
      DeleteSearchStateData((WeightedAStarState*)state->PlannerSpecificData);
      free((WeightedAStarState*)state->PlannerSpecificData);
      state->PlannerSpecificData = NULL;
    }
        }
        pSearchStateSpace->searchMDP.Delete();
}



//reset properly search state space
//needs to be done before deleting states
int WeightedAStar::ResetSearchStateSpace(WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
        pSearchStateSpace->heap->makeemptyheap();
        pSearchStateSpace->inconslist->makeemptylist(WeightedAStar_INCONS_LIST_ID);

        return 1;
}

//initialization before each search
void WeightedAStar::ReInitializeSearchStateSpace(WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
        CKey key;

        //increase callnumber
        pSearchStateSpace->callnumber++;

        //reset iteration
        pSearchStateSpace->searchiteration = 0;
        pSearchStateSpace->bNewSearchIteration = true;

#if DEBUG
    SBPL_FPRINTF(fDeb, "reinitializing search state-space (new call number=%d search iter=%d)\n", 
            pSearchStateSpace->callnumber,pSearchStateSpace->searchiteration );
#endif



        pSearchStateSpace->heap->makeemptyheap();
        pSearchStateSpace->inconslist->makeemptylist(WeightedAStar_INCONS_LIST_ID);

    //reset 
        pSearchStateSpace->eps = this->finitial_eps;
    pSearchStateSpace->eps_satisfied = INFINITECOST;

        //initialize start state
        WeightedAStarState* startstateinfo = (WeightedAStarState*)(pSearchStateSpace->searchstartstate->PlannerSpecificData);
        if(startstateinfo->callnumberaccessed != pSearchStateSpace->callnumber)
                ReInitializeSearchStateInfo(startstateinfo, pSearchStateSpace);

        startstateinfo->g = 0;

        //insert start state into the heap
        key.key[0] = (long int)(pSearchStateSpace->eps*startstateinfo->h);
        //key.key[1] = startstateinfo->h;
        pSearchStateSpace->heap->insertheap(startstateinfo, key);

    pSearchStateSpace->bReinitializeSearchStateSpace = false;
        pSearchStateSpace->bReevaluatefvals = false;
}

//very first initialization
int WeightedAStar::InitializeSearchStateSpace(WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{

        if(pSearchStateSpace->heap->currentsize != 0 || 
                pSearchStateSpace->inconslist->currentsize != 0)
        {
                SBPL_ERROR("ERROR in InitializeSearchStateSpace: heap or list is not empty\n");
                throw new SBPL_Exception();
        }

        pSearchStateSpace->eps = this->finitial_eps;
    pSearchStateSpace->eps_satisfied = INFINITECOST;
        pSearchStateSpace->searchiteration = 0;
        pSearchStateSpace->bNewSearchIteration = true;
        pSearchStateSpace->callnumber = 0;
        pSearchStateSpace->bReevaluatefvals = false;


        //create and set the search start state
        pSearchStateSpace->searchgoalstate = NULL;
        //pSearchStateSpace->searchstartstate = GetState(SearchStartStateID, pSearchStateSpace);
    pSearchStateSpace->searchstartstate = NULL;
        

    pSearchStateSpace->bReinitializeSearchStateSpace = true;

        return 1;

}


int WeightedAStar::SetSearchGoalState(int SearchGoalStateID, WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{
        if(pSearchStateSpace->searchgoalstate == NULL || 
                pSearchStateSpace->searchgoalstate->StateID != SearchGoalStateID)
        {
                pSearchStateSpace->searchgoalstate = GetState(SearchGoalStateID, pSearchStateSpace);

                //should be new search iteration
                pSearchStateSpace->eps_satisfied = INFINITECOST;
                pSearchStateSpace->bNewSearchIteration = true;
                pSearchStateSpace_->eps = this->finitial_eps;


                //recompute heuristic for the heap if heuristics is used
#if USE_HEUR
                for(int i = 0; i < (int)pSearchStateSpace->searchMDP.StateArray.size(); i++)
                {
                        CMDPSTATE* MDPstate = pSearchStateSpace->searchMDP.StateArray[i];
                        WeightedAStarState* state = (WeightedAStarState*)MDPstate->PlannerSpecificData;
                        state->h = ComputeHeuristic(MDPstate, pSearchStateSpace);
                }
                
                pSearchStateSpace->bReevaluatefvals = true;
#endif
        }


        return 1;

}


int WeightedAStar::SetSearchStartState(int SearchStartStateID, WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{

        CMDPSTATE* MDPstate = GetState(SearchStartStateID, pSearchStateSpace);

        if(MDPstate !=  pSearchStateSpace->searchstartstate)
        {       
                pSearchStateSpace->searchstartstate = MDPstate;
                pSearchStateSpace->bReinitializeSearchStateSpace = true;
        }

        return 1;

}



int WeightedAStar::ReconstructPath(WeightedAStarSearchStateSpace_t* pSearchStateSpace)
{       


        if(bforwardsearch) //nothing to do, if search is backward
        {
                CMDPSTATE* MDPstate = pSearchStateSpace->searchgoalstate;
                CMDPSTATE* PredMDPstate;
                WeightedAStarState *predstateinfo, *stateinfo;



#if DEBUG
                SBPL_FPRINTF(fDeb, "reconstructing a path:\n");
#endif

                while(MDPstate != pSearchStateSpace->searchstartstate)
                {
                        stateinfo = (WeightedAStarState*)MDPstate->PlannerSpecificData;

#if DEBUG
                        PrintSearchState(stateinfo, fDeb);
#endif
                        if(stateinfo->g == INFINITECOST)
                        {       
                                //SBPL_ERROR("ERROR in ReconstructPath: g of the state on the path is INFINITE\n");
                                //throw new SBPL_Exception();
                                return -1;
                        }

                        if(stateinfo->bestpredstate == NULL)
                        {
                                SBPL_ERROR("ERROR in ReconstructPath: bestpred is NULL\n");
                                throw new SBPL_Exception();
                        }

                        //get the parent state
                        PredMDPstate = stateinfo->bestpredstate;
                        predstateinfo = (WeightedAStarState*)PredMDPstate->PlannerSpecificData;

                        //set its best next info
                        predstateinfo->bestnextstate = MDPstate;

                        //check the decrease of g-values along the path
                        if(predstateinfo->v >= stateinfo->g)
                        {
                                SBPL_ERROR("ERROR in ReconstructPath: g-values are non-decreasing\n");                  
                                PrintSearchState(predstateinfo, fDeb);
                                throw new SBPL_Exception();
                        }

                        //transition back
                        MDPstate = PredMDPstate;
                }
        }

        return 1;
}



void WeightedAStar::PrintSearchPath(WeightedAStarSearchStateSpace_t* pSearchStateSpace, FILE* fOut)
{
        WeightedAStarState* searchstateinfo;
        CMDPSTATE* state;
        int goalID;
        int PathCost;

        if(bforwardsearch)
        {
                state  = pSearchStateSpace->searchstartstate;
                goalID = pSearchStateSpace->searchgoalstate->StateID;
        }
        else
        {
                state = pSearchStateSpace->searchgoalstate;
                goalID = pSearchStateSpace->searchstartstate->StateID;
        }
        if(fOut == NULL)
                fOut = stdout;

        PathCost = ((WeightedAStarState*)pSearchStateSpace->searchgoalstate->PlannerSpecificData)->g;

        SBPL_FPRINTF(fOut, "Printing a path from state %d to the goal state %d\n", 
                        state->StateID, pSearchStateSpace->searchgoalstate->StateID);
        SBPL_FPRINTF(fOut, "Path cost = %d:\n", PathCost);
                        
        
        environment_->PrintState(state->StateID, false, fOut);

        int costFromStart = 0;
        while(state->StateID != goalID)
        {
                SBPL_FPRINTF(fOut, "state %d ", state->StateID);

                if(state->PlannerSpecificData == NULL)
                {
                        SBPL_FPRINTF(fOut, "path does not exist since search data does not exist\n");
                        break;
                }

                searchstateinfo = (WeightedAStarState*)state->PlannerSpecificData;

                if(searchstateinfo->bestnextstate == NULL)
                {
                        SBPL_FPRINTF(fOut, "path does not exist since bestnextstate == NULL\n");
                        break;
                }
                if(searchstateinfo->g == INFINITECOST)
                {
                        SBPL_FPRINTF(fOut, "path does not exist since bestnextstate == NULL\n");
                        break;
                }

                int costToGoal = PathCost - costFromStart;
                int transcost = searchstateinfo->g - ((WeightedAStarState*)(searchstateinfo->bestnextstate->PlannerSpecificData))->v;
                if(bforwardsearch)
                        transcost = -transcost;

                costFromStart += transcost;

                SBPL_FPRINTF(fOut, "g=%d-->state %d, h = %d ctg = %d  ", searchstateinfo->g,                    
                        searchstateinfo->bestnextstate->StateID, searchstateinfo->h, costToGoal);

                state = searchstateinfo->bestnextstate;

                environment_->PrintState(state->StateID, false, fOut);



        }
}

void WeightedAStar::PrintSearchState(WeightedAStarState* state, FILE* fOut)
{
        SBPL_FPRINTF(fOut, "state %d: h=%d g=%u v=%u iterc=%d callnuma=%d expands=%d heapind=%d inconslist=%d\n",
                state->MDPstate->StateID, state->h, state->g, state->v, 
                state->iterationclosed, state->callnumberaccessed, state->numofexpands, state->heapindex, state->listelem[WeightedAStar_INCONS_LIST_ID]?1:0);
        environment_->PrintState(state->MDPstate->StateID, true, fOut);

}



int WeightedAStar::getHeurValue(WeightedAStarSearchStateSpace_t* pSearchStateSpace, int StateID)
{
        CMDPSTATE* MDPstate = GetState(StateID, pSearchStateSpace);
        WeightedAStarState* searchstateinfo = (WeightedAStarState*)MDPstate->PlannerSpecificData;
        return searchstateinfo->h;
}


vector<int> WeightedAStar::GetSearchPath(WeightedAStarSearchStateSpace_t* pSearchStateSpace, int& solcost)
{
  vector<int> SuccIDV;
  vector<int> CostV;
  vector<int> wholePathIds;
  WeightedAStarState* searchstateinfo;
  CMDPSTATE* state = NULL; 
  CMDPSTATE* goalstate = NULL;
  CMDPSTATE* startstate=NULL;
  
  if(bforwardsearch){   
    startstate = pSearchStateSpace->searchstartstate;
    goalstate = pSearchStateSpace->searchgoalstate;
    
    //reconstruct the path by setting bestnextstate pointers appropriately
    ReconstructPath(pSearchStateSpace);
  }
  else{
    startstate = pSearchStateSpace->searchgoalstate;
    goalstate = pSearchStateSpace->searchstartstate;
  }
  
  
  state = startstate;
  
  wholePathIds.push_back(state->StateID);
  solcost = 0;
  
  FILE* fOut = stdout;
  while(state->StateID != goalstate->StateID){
    if(state->PlannerSpecificData == NULL){
      SBPL_FPRINTF(fOut, "path does not exist since search data does not exist\n");
      break;
    }
      
    searchstateinfo = (WeightedAStarState*)state->PlannerSpecificData;
    
    if(searchstateinfo->bestnextstate == NULL){
      SBPL_FPRINTF(fOut, "path does not exist since bestnextstate == NULL\n");
      break;
    }
    if(searchstateinfo->g == INFINITECOST){
      SBPL_FPRINTF(fOut, "path does not exist since bestnextstate == NULL\n");
      break;
    }
      
    environment_->GetSuccs(state->StateID, &SuccIDV, &CostV);
    int actioncost = INFINITECOST;
    for(int i = 0; i < (int)SuccIDV.size(); i++){   
      if(SuccIDV.at(i) == searchstateinfo->bestnextstate->StateID && CostV.at(i)<actioncost)
        actioncost = CostV.at(i);
    }
    if(actioncost == INFINITECOST)
      SBPL_PRINTF("WARNING: actioncost = %d (%d)\n", actioncost,state->StateID);
      
    solcost += actioncost;
      
    //SBPL_FPRINTF(fDeb, "actioncost=%d between states %d and %d\n", 
    //        actioncost, state->StateID, searchstateinfo->bestnextstate->StateID);
    //environment_->PrintState(state->StateID, false, fDeb);
    //environment_->PrintState(searchstateinfo->bestnextstate->StateID, false, fDeb);
      
      
#if DEBUG
    WeightedAStarState* nextstateinfo = (WeightedAStarState*)(searchstateinfo->bestnextstate->PlannerSpecificData);
    if(actioncost != abs((int)(searchstateinfo->g - nextstateinfo->g)) && pSearchStateSpace->eps_satisfied <= 1.001){
      SBPL_FPRINTF(fDeb, "ERROR: actioncost=%d is not matching the difference in g-values of %d\n", 
        actioncost, abs((int)(searchstateinfo->g - nextstateinfo->g)));
      SBPL_ERROR("ERROR: actioncost=%d is not matching the difference in g-values of %d\n", 
       actioncost,abs((int)(searchstateinfo->g - nextstateinfo->g)));
      PrintSearchState(searchstateinfo, fDeb);
      PrintSearchState(nextstateinfo, fDeb);
    }
#endif
      
      
    state = searchstateinfo->bestnextstate;
      
    wholePathIds.push_back(state->StateID);
  }


  return wholePathIds;
}



bool WeightedAStar::Search(WeightedAStarSearchStateSpace_t* pSearchStateSpace, vector<int>& pathIds, int & PathCost, bool bFirstSolution, bool bOptimalSolution, double MaxNumofSecs)
{
        CKey key;
        TimeStarted = clock();
    searchexpands = 0;

#if DEBUG
        SBPL_FPRINTF(fDeb, "new search call (call number=%d)\n", pSearchStateSpace->callnumber);
#endif

    if(pSearchStateSpace->bReinitializeSearchStateSpace == true){
        //re-initialize state space 
        ReInitializeSearchStateSpace(pSearchStateSpace);
    }


        if(bOptimalSolution)
        {
                pSearchStateSpace->eps = 1;
                MaxNumofSecs = INFINITECOST;
        }
        else if(bFirstSolution)
        {
                MaxNumofSecs = INFINITECOST;
        }

        //the main loop of WeightedAStar*
        int prevexpands = 0;
        clock_t loop_time;
        while(pSearchStateSpace->eps_satisfied > WeightedAStar_FINAL_EPS && 
                (clock()- TimeStarted) < MaxNumofSecs*(double)CLOCKS_PER_SEC)
        {
        loop_time = clock();
                //decrease eps for all subsequent iterations
                if(fabs(pSearchStateSpace->eps_satisfied - pSearchStateSpace->eps) < ERR_EPS && !bFirstSolution)
                {
                        pSearchStateSpace->eps = pSearchStateSpace->eps - WeightedAStar_DECREASE_EPS;
                        if(pSearchStateSpace->eps < WeightedAStar_FINAL_EPS)
                                pSearchStateSpace->eps = WeightedAStar_FINAL_EPS;

                        //the priorities need to be updated
                        pSearchStateSpace->bReevaluatefvals = true; 

                        //it will be a new search
                        pSearchStateSpace->bNewSearchIteration = true;

                        //build a new open list by merging it with incons one
                        BuildNewOPENList(pSearchStateSpace); 

                }

                if(pSearchStateSpace->bNewSearchIteration)
                {
                        pSearchStateSpace->searchiteration++;
                        pSearchStateSpace->bNewSearchIteration = false;
                }

                //re-compute f-values if necessary and reorder the heap
                if(pSearchStateSpace->bReevaluatefvals) 
                        Reevaluatefvals(pSearchStateSpace);

                //improve or compute path
                if(ImprovePath(pSearchStateSpace, MaxNumofSecs) == 1){
            pSearchStateSpace->eps_satisfied = pSearchStateSpace->eps;
        }

                //print the solution cost and eps bound
                SBPL_PRINTF("eps=%f expands=%d g(searchgoal)=%d time=%.3f\n", pSearchStateSpace->eps_satisfied, searchexpands - prevexpands,
                                                        ((WeightedAStarState*)pSearchStateSpace->searchgoalstate->PlannerSpecificData)->g,double(clock()-loop_time)/CLOCKS_PER_SEC);

#if DEBUG
        SBPL_FPRINTF(fDeb, "eps=%f expands=%d g(searchgoal)=%d time=%.3f\n", pSearchStateSpace->eps_satisfied, searchexpands - prevexpands,
                                                        ((WeightedAStarState*)pSearchStateSpace->searchgoalstate->PlannerSpecificData)->g,double(clock()-loop_time)/CLOCKS_PER_SEC);
                PrintSearchState((WeightedAStarState*)pSearchStateSpace->searchgoalstate->PlannerSpecificData, fDeb);
#endif
                prevexpands = searchexpands;


                //if just the first solution then we are done
                if(bFirstSolution)
                        break;

                //no solution exists
                if(((WeightedAStarState*)pSearchStateSpace->searchgoalstate->PlannerSpecificData)->g == INFINITECOST)
                        break;

        }


#if DEBUG
        fflush(fDeb);
#endif

        PathCost = ((WeightedAStarState*)pSearchStateSpace->searchgoalstate->PlannerSpecificData)->g;
        MaxMemoryCounter += environment_->StateID2IndexMapping.size()*sizeof(int);
        
        SBPL_PRINTF("MaxMemoryCounter = %d\n", MaxMemoryCounter);

        int solcost = INFINITECOST;
        bool ret = false;
        if(PathCost == INFINITECOST)
        {
                SBPL_PRINTF("could not find a solution\n");
                ret = false;
        }
        else
        {
                SBPL_PRINTF("solution is found\n");      
        pathIds = GetSearchPath(pSearchStateSpace, solcost);
        ret = true;
        }

  solfound = ret;
  searchtime = (clock()-TimeStarted)/((double)CLOCKS_PER_SEC);
  finalsolcost = solcost;
        SBPL_PRINTF("total expands this call = %d, planning time = %.3f secs, solution cost=%d\n", 
           searchexpands, searchtime, solcost);
    

    //SBPL_FPRINTF(fStat, "%d %d\n", searchexpands, solcost);

        return ret;

}


//-----------------------------Interface function-----------------------------------------------------
//returns 1 if found a solution, and 0 otherwise
int WeightedAStar::replan(double allocated_time_secs, vector<int>* solution_stateIDs_V)
{
        int solcost;

        return replan(allocated_time_secs, solution_stateIDs_V, &solcost);
        
}

//returns 1 if found a solution, and 0 otherwise
int WeightedAStar::replan(double allocated_time_secs, vector<int>* solution_stateIDs_V, int* psolcost)
{
  bubbles.resize(environment_->getNumBubbles());
  for(unsigned int i=0; i<bubbles.size(); i++)
    bubbles[i].active = false;

  vector<int> pathIds; 
  bool bFound = false;
  int PathCost;
  bool bFirstSolution = this->bsearchuntilfirstsolution;
  bool bOptimalSolution = false;
  *psolcost = 0;
  
  SBPL_PRINTF("planner: replan called (bFirstSol=%d, bOptSol=%d)\n", bFirstSolution, bOptimalSolution);
  
  //plan
  if((bFound = Search(pSearchStateSpace_, pathIds, PathCost, bFirstSolution, bOptimalSolution, allocated_time_secs)) == false) 
    {
      SBPL_PRINTF("failed to find a solution\n");
    }
  
  //copy the solution
  *solution_stateIDs_V = pathIds;
  *psolcost = PathCost;

  //environment_->PostProcess(solution_stateIDs_V, psolcost);
  
  return (int)bFound;

}


int WeightedAStar::set_goal(int goal_stateID)
{

        SBPL_PRINTF("planner: setting goal to %d\n", goal_stateID);
        environment_->PrintState(goal_stateID, true, stdout);

        if(bforwardsearch)
        {       
                if(SetSearchGoalState(goal_stateID, pSearchStateSpace_) != 1)
                        {
                                SBPL_ERROR("ERROR: failed to set search goal state\n");
                                return 0;
                        }
        }
        else
        {
            if(SetSearchStartState(goal_stateID, pSearchStateSpace_) != 1)
        {
            SBPL_ERROR("ERROR: failed to set search start state\n");
            return 0;
        }
        }

    return 1;
}


int WeightedAStar::set_start(int start_stateID)
{

        SBPL_PRINTF("planner: setting start to %d\n", start_stateID);
        environment_->PrintState(start_stateID, true, stdout);

        if(bforwardsearch)
        {       

            if(SetSearchStartState(start_stateID, pSearchStateSpace_) != 1)
        {
            SBPL_ERROR("ERROR: failed to set search start state\n");
            return 0;
        }
        }
        else
        {
            if(SetSearchGoalState(start_stateID, pSearchStateSpace_) != 1)
        {
            SBPL_ERROR("ERROR: failed to set search goal state\n");
            return 0;
        }
        }

    return 1;

}



void WeightedAStar::costs_changed(StateChangeQuery const & stateChange)
{


    pSearchStateSpace_->bReinitializeSearchStateSpace = true;


}

void WeightedAStar::costs_changed()
{

    pSearchStateSpace_->bReinitializeSearchStateSpace = true;

}



int WeightedAStar::force_planning_from_scratch()
{
        SBPL_PRINTF("planner: forceplanfromscratch set\n");

    pSearchStateSpace_->bReinitializeSearchStateSpace = true;

    return 1;
}


int WeightedAStar::set_search_mode(bool bSearchUntilFirstSolution)
{

        SBPL_PRINTF("planner: search mode set to %d\n", bSearchUntilFirstSolution);

        bsearchuntilfirstsolution = bSearchUntilFirstSolution;

        return 1;
}


void WeightedAStar::print_searchpath(FILE* fOut)
{
        PrintSearchPath(pSearchStateSpace_, fOut);
}


//---------------------------------------------------------------------------------------------------------

//function for getting search data
void WeightedAStar::getSearchStats(bool* solutionFound, int* numExpands, int* solutionCost, double* searchTime){
  *solutionFound = solfound;
  *numExpands = searchexpands;
  *solutionCost = finalsolcost;
  *searchTime = searchtime;
}

---

sbpl_dynamic_planner
Author(s): Michael Phillips, Maxim Likhachev
autogenerated on Fri Jan 11 09:41:03 2013