games/vcmi
1
0
Fork 0
mirror of https://github.com/vcmi/vcmi.git synced 2025-03-19 21:10:12 +02:00
Code Issues Releases Activity

[programming challenge, SSN]

Browse Source 
* inpout scaling
* function for evaluating artifact for given battle
* additional input for ANN
* increased hidden layer size
* still very, *very* slow
This commit is contained in:
mateuszb 2012-05-26 13:27:32 +00:00
parent a281483978
commit 4ae3c8c8f1
1 changed files with 31 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

48
Odpalarka/main.cpp
View File

@ -119,8 +119,8 @@ std::vector<CArtifactInstance*> genArts(const std::vector<Bonus> & bonusesToGive
return ret;
}
//returns how good the artifact is for the neural network
double runSSN(FANN::neural_net & net, const DuelParameters dp, CArtifactInstance * inst)
//rates given artifact
double rateArt(const DuelParameters dp, CArtifactInstance * inst)
{
TArtSet setL, setR;
setL[inst->artType->possibleSlots[0]] = inst;
@ -135,7 +135,7 @@ double runSSN(FANN::neural_net & net, const DuelParameters dp, CArtifactInstance
}
const unsigned int num_input = 24;
const unsigned int num_input = 27;
double * genSSNinput(const DuelParameters & dp, CArtifactInstance * art)
{
@ -144,17 +144,17 @@ double * genSSNinput(const DuelParameters & dp, CArtifactInstance * art)
//general description
*(cur++) = dp.bfieldType;
*(cur++) = dp.terType;
*(cur++) = dp.bfieldType/30.0;
*(cur++) = dp.terType/12.0;
//creature & hero description
for(int i=0; i<2; ++i)
{
auto & side = dp.sides[0];
*(cur++) = side.heroId;
*(cur++) = side.heroId/200.0;
for(int k=0; k<4; ++k)
*(cur++) = side.heroPrimSkills[k];
*(cur++) = side.heroPrimSkills[k]/20.0;
//weighted average of statistics
auto avg = [&](std::function<int(CCreature *)> getter) -> double
@ -173,19 +173,33 @@ double * genSSNinput(const DuelParameters & dp, CArtifactInstance * art)
return ret/div;
};
*(cur++) = avg([](CCreature * c){return c->attack;});
*(cur++) = avg([](CCreature * c){return c->defence;});
*(cur++) = avg([](CCreature * c){return c->speed;});
*(cur++) = avg([](CCreature * c){return c->hitPoints;});
*(cur++) = avg([](CCreature * c){return c->attack;})/50.0;
*(cur++) = avg([](CCreature * c){return c->defence;})/50.0;
*(cur++) = avg([](CCreature * c){return c->speed;})/15.0;
*(cur++) = avg([](CCreature * c){return c->hitPoints;})/1000.0;
}
//bonus description
auto & blist = art->getBonusList();
*(cur++) = art->Attack();
*(cur++) = art->Defense();
*(cur++) = blist.valOfBonuses(Selector::type(Bonus::STACKS_SPEED));
*(cur++) = blist.valOfBonuses(Selector::type(Bonus::STACK_HEALTH));
*(cur++) = blist[0]->type/100.0;
*(cur++) = blist[0]->subtype/10.0;
*(cur++) = blist[0]->val/100.0;;
*(cur++) = art->Attack()/10.0;
*(cur++) = art->Defense()/10.0;
*(cur++) = blist.valOfBonuses(Selector::type(Bonus::STACKS_SPEED))/5.0;
*(cur++) = blist.valOfBonuses(Selector::type(Bonus::STACK_HEALTH))/10.0;
return ret;
}
//returns how good the artifact is for the neural network
double runSSN(FANN::neural_net & net, const DuelParameters dp, CArtifactInstance * inst)
{
double * input = genSSNinput(dp, inst);
double * out = net.run(input);
double ret = *out;
free(out);
return ret;
}
@ -200,7 +214,7 @@ void learnSSN(FANN::neural_net & net, const std::vector<std::pair<DuelParameters
{
inputs[i] = genSSNinput(input[i].first, input[i].second);
outputs[i] = new double;
*(outputs[i]) = runSSN(net, input[i].first, input[i].second);
*(outputs[i]) = rateArt(input[i].first, input[i].second);
}
td.set_train_data(input.size(), num_input, inputs, 1, outputs);
@ -211,7 +225,7 @@ void initNet(FANN::neural_net & ret)
{
const float learning_rate = 0.7f;
const unsigned int num_layers = 3;
const unsigned int num_hidden = 3;
const unsigned int num_hidden = 30;
const unsigned int num_output = 1;
const float desired_error = 0.001f;
const unsigned int max_iterations = 300000;