int main(int argc, const char * argv[])
{
using namespace distributed_solver;
// Add scenarios and path to store instances here...
//int advertiser_dimensions [] = {1000};
//int impression_dimensions [] = {1000};
//long double bid_sparsity_scenario [] = {0.1};
//int advertiser_dimensions [] = {2};
//int impression_dimensions [] = {10};
//long double bid_sparsity_scenario [] = {0.5};
int advertiser_dimensions [] = {10000};
int impression_dimensions [] = {10000};
long double bid_sparsity_scenario [] = {0.001};
std::string file_name_path = "/Users/ciocan/Documents/Google/data/experiment_";
int num_iterations = 1000;
long double epsilon = 0.05;
long double numerical_accuracy_tolerance = 0.000000000000000001;
for (int i = 0; i < (sizeof(advertiser_dimensions) / sizeof(int)); ++i) {
for (int j = 0; j < (sizeof(impression_dimensions) / sizeof(int)); ++j) {
for (int k = 0; k < (sizeof(bid_sparsity_scenario) / sizeof(long double)); ++k) {
Instance inst = Instance(advertiser_dimensions[i],
impression_dimensions[j],
1,
bid_sparsity_scenario[k],
epsilon,
0.25,
numerical_accuracy_tolerance);
inst.GenerateInstance();
// std::cout << file_name_path + "\n";
// inst.WriteInstanceToCSV(file_name_path);
// inst.GenerateAndWriteInstance(file_name_path);
// std::cout << "created global problem \n";
inst.RunMultiplicativeWeights(num_iterations, numerical_accuracy_tolerance);
std::cout << "finished \n";
}
}
}
return 0;
}
int main(int argc, const char * argv[])
{
clock_t t1 = clock();
using namespace distributed_solver;
// Add scenarios and path to store instances here...
int advertiser_dimensions [] = {1000};
int impression_dimensions [] = {1000};
long double bid_sparsity_scenario [] = {0.1};
//int advertiser_dimensions [] = {2};
//int impression_dimensions [] = {10};
//long double bid_sparsity_scenario [] = {0.5};
//int advertiser_dimensions [] = {100000};
//int impression_dimensions [] = {1000000};
//long double bid_sparsity_scenario [] = {0.0001};
std::string file_name_path = "/Users/ciocan/Documents/Google/data/experiment_";
int num_iterations = 300;
long double epsilon = 0.01;
long double numerical_accuracy_tolerance = 0.000000000000000001;
bool use_binary_search = true; // Setting this to false runs the sort method
int num_bin_intervals = 3; // Sets the number of critical ratios to calculate
long double cr_transition_scale = 1 - epsilon * 0.001;
for (int i = 0; i < (sizeof(advertiser_dimensions) / sizeof(int)); ++i) {
for (int j = 0; j < (sizeof(impression_dimensions) / sizeof(int)); ++j) {
for (int k = 0; k < (sizeof(bid_sparsity_scenario) / sizeof(long double)); ++k) {
Instance inst = Instance(advertiser_dimensions[i],
impression_dimensions[j],
1,
bid_sparsity_scenario[k],
epsilon,
0.25,
numerical_accuracy_tolerance);
inst.GenerateInstance();
// std::cout << file_name_path + "\n";
// inst.WriteInstanceToCSV(file_name_path);
// inst.GenerateAndWriteInstance(file_name_path);
// std::cout << "created global problem \n";
if(!use_binary_search){
inst.RunMultiplicativeWeights(num_iterations, numerical_accuracy_tolerance, use_binary_search);
}
else{
inst.RunMultiplicativeWeights(num_iterations, numerical_accuracy_tolerance, use_binary_search,
cr_transition_scale, num_bin_intervals);
}
std::cout << "finished \n";
}
}
}
clock_t t2 = clock();
cout << "Total running time: " << (t2-t1) << endl;
return 0;
}