/// <summary>
/// Randomly generates a firm object (production technology and output market parameters).
/// </summary>
/// <param name="ip">A pointer to the collection of input parameters.</param>
/// <param name="FirmID">Unique identifier for this firm (run number)</param>
public Firm(InputParameters ip, int FirmID)
{
// Choose random values for DISP2 (the top DISP1 resources
// account for DISP2 percent of total resource costs), and
// density (sparsity) of resource consumption pattern matrix
this.g = GenRandNumbers.GenUniformDbl(ip.DISP2_MIN, ip.DISP2_MAX);
this.d = GenRandNumbers.GenUniformDbl(ip.DNS_MIN, ip.DNS_MAX);
// Generate the true product margins and the true, optimal
// decision vector. Keep generating new margins until there
// is at least one product in the optimal mix.
RowVector MAR, DECT0;
do
{
MAR = this.GenMargins(ip);
DECT0 = MAR.Map(x => (x < 1.0) ? 0.0 : 1.0);
} while (DECT0.TrueForAll(x => x == 0.0));
// Generate vector of maximum production quantities
this.mxq = this.GenMXQ(ip);
// And associated vector of optimal production quantities
ColumnVector QT = mxq.ewMultiply(DECT0);
// Flowchart 5.1 - Create resource consumption pattern matrix
this.res_cons_pat = GenResConsPat(ip);
// Flowchart 5.2 - Compute TRU
// Calculate vector of total units of resource
// consumption, by product
ColumnVector TRU = this.CalcResConsumption(QT);
// Flowchart 5.3 - Compute MAXRU
// Calculate resource consumption under the assumption
// that all products are produced at maximum quantity
ColumnVector MAXRU = this.CalcResConsumption(mxq);
RowVector RCC, PC_B, RCCN;
double TCT0;
#region Flowchart 5.4 - Generate RCC, RCU, and RCCN
/* -------------------------------- */
// Flowchart 5.4(a)-(g)
// Generate vector of total resource costs (RCC)
RCC = GenRCC(ip);
/* -------------------------------- */
// Flowchart 5.4(h)
// Now generate unit resource costs (RCU) by doing element-wise
// division of RCC by MAXRU
this.rcu = RCC.Map((x, i) => x / MAXRU[i]);
/* -------------------------------- */
// Flowchart 5.4(i)
// Compute new RCC vector (RCCN) based on unit resource
// costs (RCU) and true unit resource consumption (TRU)
RCCN = this.rcu.ewMultiply(TRU);
// Check to see if the first resource (RCCN[0]) is the largest.
// If not, increase RCU[0] by just enough to make it so.
if (RCCN[0] < RCCN.Skip(1).Max() + 1)
{
RCCN[0] = Math.Ceiling(RCCN.Max()) + 1.0;
this.rcu[0] = RCCN[0] / TRU[0];
}
#endregion
// Flowchart 5.5 - Calculate PC_B
// Calculate true unit product costs
PC_B = this.CalcTrueProductCosts();
// Flowchart 5.6 - Compute total costs TCT0
// Compute total costs
TCT0 = this.CalcTotCosts(QT);
// Flowchart 5.7 - Rename RCCN to RCC
RCC = RCCN;
initial_rcc = RCC;
#region Flowchart 5.8 - Calculate SP, TRV0, PROFITT0
// Calculate product selling prices, total revenue, and profit
this.sp = PC_B.ewMultiply(MAR);
double TRV0 = this.sp * QT;
this.profitt0 = TRV0 - TCT0;
#endregion
// 5.9(a) Create RANK vector
// Note: this method provides a stable sort. It's important to use a stable sort.
// LOOKUP IN VERSION.TXT WHY IT'S IMPORTANT TO USE A STABLE SORT HERE.
initial_rank = Enumerable.Range(0, RCC.Dimension).OrderByDescending(i => RCC[i]).ToArray();
#region Flowchart 5.9(b) - Create RES_CONS_PAT_PRCT
this.res_cons_pat_prct = new RectangularMatrix(ip.RCP, ip.CO);
for (int r = 0; r < this.res_cons_pat.RowCount; ++r)
{
RowVector rv = this.res_cons_pat.Row(r);
if (TRU[r] != 0.0)
{
rv = rv.Map((alt_ij, col) => alt_ij * QT[col] / TRU[r]);
if (Math.Abs(rv.Sum() - 1.0) > 0.01)
{
throw new ApplicationException("Sum of row of RES_CONS_PAT_PRCT not equal to 1.");
}
}
else
{
rv = rv.Map(alt_ij => 0.0);
}
this.res_cons_pat_prct.CopyRowInto(rv, r);
}
#endregion
#region Flowchart 5.9(c) - Create correlation matrix
// Create correlation matrix for rows of RES_CONS_PAT_PRCT
MultivariateSample mvs = new MultivariateSample(ip.RCP);
for (int c = 0; c < this.res_cons_pat_prct.ColumnCount; ++c)
{
mvs.Add(this.res_cons_pat_prct.Column(c));
}
this.pearsoncorr = new SymmetricMatrix(ip.RCP);
for (int i = 0; i < mvs.Dimension; ++i)
{
for (int j = i; j < mvs.Dimension; ++j)
{
//PearsonCorr[i, j] = mvs.PearsonRTest( i, j ).Statistic;
this.pearsoncorr[i, j] = mvs.TwoColumns(i, j).PearsonRTest().Statistic;
}
}
#endregion
// Flowchart 5.10 - Logging true system
// Note: I'm deliberately passing copies of the fields MXQ, SP, etc.
Output.LogFirm(
ip, this, FirmID,
MAR, DECT0,
TRV0, TCT0, profitt0,
RCC);
}
/// <summary>
/// Implements step 5.3 of the flowchart: Generates a [1 x RCP] vector of
/// total resource costs by resource.
/// </summary>
/// <param name="ip">An input parameters object.</param>
private RowVector GenRCC(InputParameters ip)
{
bool throwAway;
int numThrows = 0;
List <double> rcc;
// repeat the following loop until a suitable vector
// RCC is generated.
do
{
/* -------------------------------- */
// Flowchart 5.4(b)
// Calculate total resource cost of first DISP1 resources
double topTR = G * ip.TR;
// Calculate minimum allowable resource cost in first
// DISP1 resources
double rmin = (1.0 - G) * ip.TR / (ip.RCP - ip.DISP1);
// The following is an upward adjustment of rmin.
// Without this, the values of the resources in the
// remaining resources have too little variance.
// It checks how much room there is to adjust rmin,
// and takes 2.5% of that room. The 2.5% was determined
// through trial and error.
double maxValOfLargestElem = topTR - ((ip.DISP1 - 1.0) * rmin);
rmin += (maxValOfLargestElem - rmin) * 0.025;
/* -------------------------------- */
// Flowchart 5.4(c)
// Generate the first DISP1 random numbers
List <double> temp1 = new List <double>();
for (int i = 1; i <= ip.DISP1 - 1; ++i)
{
double rmax = (topTR - temp1.Sum()) - ((ip.DISP1 - i) * rmin);
if (rmax < rmin)
{
throw new ApplicationException("rmax less than rmin");
}
double ri = GenRandNumbers.GenUniformDbl(rmin, rmax);
temp1.Add(ri);
}
// The final element is computed to ensure that the total
// in temp1 is topTR
temp1.Add(topTR - temp1.Sum());
// Move the biggest resource to the front
double temp1Max = temp1.Max();
if (!temp1.Remove(temp1Max))
{
throw new ApplicationException("Could not remove largest element.");
}
temp1.Insert(0, temp1Max);
// SOME CHECKS ON THE NUMBERS
if (Math.Abs(temp1.Sum() - ip.TR * G) > 1.0)
{
throw new ApplicationException("Sum of first DISP1 resources not correct.");
}
if (temp1.Min() < (1 - G) * ip.TR / (ip.RCP - ip.DISP1))
{
throw new ApplicationException("Min element too small.");
}
/* -------------------------------- */
// Flowchart 5.4(d)
List <double> temp2 = new List <double>();
for (int i = 0; i < ip.RCP - ip.DISP1; ++i)
{
temp2.Add(GenRandNumbers.GenUniformDbl(0.05, 0.95));
}
temp2.Normalize();
temp2.MultiplyBy((1.0 - G) * ip.TR);
double temp1Min = temp1.Min();
while (temp2.Max() - temp1.Min() > 1.0)
{
// Sort the list in descending order
temp2.Sort();
temp2.Reverse();
for (int i = 0; i < temp2.Count / 2; ++i)
{
double overage = Math.Max(temp2[i] - temp1Min, 0.0);
temp2[i] -= overage;
temp2[temp2.Count - 1 - i] += overage;
}
}
temp2.Shuffle();
// SOME CHECKS
if (Math.Abs(temp2.Sum() - ip.TR * (1.0 - G)) > 1.0)
{
throw new ApplicationException("Sum of small resources not correct.");
}
/* -------------------------------- */
// Flowchart 5.4(e)
rcc = new List <double>(ip.RCP);
rcc.AddRange(temp1);
rcc.AddRange(temp2);
/* -------------------------------- */
// Flowchart 5.4(f)
throwAway = rcc.Exists(x => x < 1.0);
// SOME CHECKS
if (rcc.Min() < 0.0)
{
throw new ApplicationException("Negative element in RCC.");
}
if (throwAway)
{
++numThrows;
}
} while (throwAway);
return(new RowVector(rcc));
}
/// <summary>
/// Generates a resource consumption pattern matrix
/// </summary>
/// <param name="ip">The current InputParameters object</param>
private RectangularMatrix GenResConsPat(InputParameters ip)
{
bool throwAway;
int numThrows = 0;
RectangularMatrix outputMatrix;
do
{
throwAway = false;
outputMatrix = new RectangularMatrix(ip.RCP, ip.CO);
// Flowchart 5.1(a): Generate vector X
RowVector X = GenRandNumbers.GenStdNormalVec(ip.CO);
// The following code is used in both 5.1(b) and 5.1(c):
RowVector[] Y = new RowVector[ip.RCP - 1];
RowVector[] Z = new RowVector[Y.Length];
for (int i = 0; i < Y.Length; ++i)
{
Y[i] = GenRandNumbers.GenStdNormalVec(ip.CO);
}
// Flowchart 5.1(b): Generate (DISP1 - 1) vectors Y
// Then create Z vectors based on X and Y
double COR1 =
GenRandNumbers.GenUniformDbl(ip.COR1LB, ip.COR1UB);
double sqrtConstant1 = Math.Sqrt(1 - COR1 * COR1);
for (int i = 0; i < ip.DISP1 - 1; ++i)
{
Z[i] = (COR1 * X) + (sqrtConstant1 * Y[i]);
}
// Flowchart 5.1(c): Generate (RCP - DISP1) vectors Y
// Then create the remaining Z vectors based on X and Y
double COR2 =
GenRandNumbers.GenUniformDbl(ip.COR2LB, ip.COR2UB);
double sqrtConstant2 = Math.Sqrt(1 - COR2 * COR2);
for (int i = ip.DISP1 - 1; i < Z.Length; ++i)
{
Z[i] = (COR2 * X) + (sqrtConstant2 * Y[i]);
}
// Flowchart 5.1(d):
// Take the absolute values of X and the Z's and
// scale both by 10.0.
X = X.Map(x => 10.0 * Math.Abs(x));
for (int i = 0; i < Z.Length; ++i)
{
Z[i] = Z[i].Map(z => 10.0 * Math.Abs(z));
}
// Round X and the Z's to integers
X = X.Map(x => Math.Ceiling(x));
for (int i = 0; i < Z.Length; ++i)
{
Z[i] = Z[i].Map(z => Math.Ceiling(z));
}
// Flowchart 5.1(e):
// Now punch out values in the Z's at random to make
// the matrix sparse
for (int i = 0; i < Z.Length; ++i)
{
Z[i] = Z[i].Map(x => ((GenRandNumbers.GenUniformDbl() < D) ? x : 0.0));
}
// Flowchart 5.1(f):
// Copy X into first row of outputMatrix.
outputMatrix.CopyRowInto(X, 0);
// Copy the Z's into the remaining rows of outputMatrix.
for (int i = 0; i < Z.Length; ++i)
{
outputMatrix.CopyRowInto(Z[i], i + 1);
}
// Ensure that the first row has no zeros
// There is a very small probability of getting a zero with
// the Ceiling function, but given that there are a finite
// number of double-precision floating point numbers, it
// is not impossible to get a 0.0.
double[] firstRow = outputMatrix.Row(0).ToArray();
if (Array.Exists(firstRow, x => x == 0.0))
{
throwAway = true;
break;
}
// Ensure that each *row* has at least one non-zero entry
for (int i = 0; i < outputMatrix.RowCount; ++i)
{
double[] nextRow = outputMatrix.Row(i).ToArray();
if (Array.TrueForAll(nextRow, x => x == 0.0))
{
throwAway = true;
break;
}
}
// Ensure that each *column* has at least one non-zero entry
// Technically, this check is redundant, as the first row, X,
// is not supposed to have any zero entries. But just to be
// on the safe side...
for (int j = 0; j < outputMatrix.ColumnCount; ++j)
{
double[] nextCol = outputMatrix.Column(j).ToArray();
if (Array.TrueForAll(nextCol, x => x == 0.0))
{
string s = "There is a column with all zeros. " +
"That should not happen since the first row is " +
"supposed to have no zeros.";
throw new ApplicationException(s);
}
}
if (throwAway)
{
++numThrows;
}
} while (throwAway);
Console.WriteLine("RES_CONS_PAT: {0} Throw aways\n", numThrows);
return(outputMatrix);
}
/// <summary>
/// Generates a vector of product margins. Each element is
/// U[ip.MARLB, ip.MARUB]. Values less than (greater) than one indicate
/// products that generate losses (profits).
/// </summary>
/// <param name="ip">The current InputParameters object</param>
/// <returns>A [1 x CO] vector, each element drawn from
/// the distribution U[ip.MARLB, ip.MARUB].</returns>
private RowVector GenMargins(InputParameters ip)
{
return(new RowVector(ip.CO)
.Map(x => GenRandNumbers.GenUniformDbl(ip.MARLB, ip.MARUB)));
}
static void Main(string[] args)
{
#region Console header
DrawASCIIart();
#endregion
#region Read input file and create InputParameters object
FileInfo inputFile = new FileInfo(Environment.CurrentDirectory + @"\input.txt");
if (!inputFile.Exists)
{
Console.WriteLine("Could not find input file: \n{0}", inputFile.FullName);
Console.WriteLine("Aborting. Press ENTER to end the program.");
Console.ReadLine();
return;
}
InputParameters ip = new InputParameters(inputFile);
#endregion
#region Make a copy of the input file
// We found it helpful to make a copy of the input file every time we ran the
// simulation. We stamp the copy's filename with the date and time so that
// we know which results files correspond to which input file.
DateTime dt = DateTime.Now;
string inputFileCopyName =
String.Format(
"input {0:D2}-{1:D2}-{2:D4} {3:D2}h {4:D2}m {5:D2}s, seed {6:G}.txt",
dt.Month,
dt.Day,
dt.Year,
dt.Hour,
dt.Minute,
dt.Second,
GenRandNumbers.GetSeed()
);
FileInfo inputFileCopy = new FileInfo(Environment.CurrentDirectory + @"\" + inputFileCopyName);
inputFile.CopyTo(inputFileCopy.FullName, true);
File.SetCreationTime(inputFileCopy.FullName, dt);
File.SetLastWriteTime(inputFileCopy.FullName, dt);
#endregion
#region Create output files
Output.CreateOutputFiles(ip);
#endregion
#region Generate Sample of Firms and their Cost Systems
Firm[] sampleFirms = new Firm[ip.NUM_FIRMS];
for (int firmID = 1; firmID <= ip.NUM_FIRMS; ++firmID)
{
Console.WriteLine(
"Starting firm {0:D3} of {1}",
firmID + 1, sampleFirms.Length
);
Firm f = new Firm(ip, firmID);
sampleFirms[firmID - 1] = f;
for (int a_indx = 0; a_indx < ip.ACP.Count; ++a_indx)
{
int a = ip.ACP[a_indx];
for (int p_indx = 0; p_indx < ip.PACP.Count; ++p_indx)
{
int p = ip.PACP[p_indx];
for (int r_indx = 0; r_indx < ip.PDR.Count; ++r_indx)
{
int r = ip.PDR[r_indx];
// Create a cost system
CostSys costsys = new CostSys(ip, f, a, p, r);
f.costSystems.Add(costsys);
int costSysID = f.costSystems.Count;
Output.LogCostSys(costsys, firmID, costSysID);
// Generate a starting decision for the cost system.
RowVector startingDecision;
if (ip.STARTMIX == 0)
{
startingDecision = f.CalcOptimalDecision();
}
else
{
var ones = Enumerable.Repeat(1.0, ip.CO).ToList();
startingDecision = new RowVector(ones);
for (int i = 0; i < startingDecision.Dimension; ++i)
{
if (GenRandNumbers.GenUniformDbl() < ip.EXCLUDE)
{
startingDecision[i] = 0.0;
}
}
}
/* Examine error in cost from implementing this decision.
* Assume the firm implements the decision startingDecision. Upon
* doing so, it will observe total resource consumption. It will then
* allocate resources to cost pools, as per the B parameter of the cost
* system, choose drivers as per the D parameter of the cost system,
* and then allocate resources to cost objects and compute reported costs.
* The reported costs are returned as PC_R. The difference
* between these and the true benchmark costs (PC_B) is used to compute
* the mean percent error in costs.
*/
RowVector PC_R = costsys.CalcReportedCosts(ip, startingDecision);
RowVector PC_B = f.CalcTrueProductCosts();
double MPE = PC_B.Zip(PC_R, (pc_b, pc_r) => Math.Abs(pc_b - pc_r) / pc_b).Sum() / PC_B.Dimension;
Output.LogCostSysError(costsys, firmID, costSysID, startingDecision, PC_B, PC_R, MPE);
/* Assume the firm implements the decision startingDecision. Upon
* doing so, it will observe total resource consumption. It will then
* allocate resources to cost pools, as per the B parameter of the cost
* system, choose drivers as per the D parameter of the cost system,
* and then allocate resources to cost objects and compute reported costs.
* The reported costs are returned as PC_R. Upon observing the
* reported costs, the firm may wish to update its original decision. When
* it implements the updated decision, costs will change again. The outcome
* of this process will either be an equilibrium decision (fixed point), or
* a cycle of decisions.
*/
(CostSystemOutcomes stopCode, RowVector endingDecision) = costsys.EquilibriumCheck(ip, startingDecision);
Output.LogCostSysLoop(costsys, firmID, costSysID, startingDecision, endingDecision, stopCode);
}
}
}
}
#endregion
Console.WriteLine("Writing output files...");
Output.WriteOutput();
Console.WriteLine("Done!");
}