Modifed to be inherited. Refs #7476.

Modify CostFuncLeastSquare such that CostFuncRwp can inherit from it.

Code/Mantid/Framework/CurveFitting/inc/MantidCurveFitting/CostFuncLeastSquares.h

@@ -94,7 +94,14 @@ protected:
     API::FunctionValues_sptr values,
     bool evalFunction = true, bool evalDeriv = true, bool evalHessian = true) const;
 
-private:
+  /// Get weight (1/sigma)
+  virtual double getWeight(API::FunctionValues_sptr values, size_t i, double sqrtW=1.0) const;
+
+  /// Calcualte sqrt(W). Final cost function = sum_i [ (obs_i - cal_i) / (sigma * sqrt(W))]**2
+  virtual double calSqrtW(API::FunctionValues_sptr values) const;
+
+  /// Flag to include constraint in cost function value
+  const bool m_includePenalty;
 
   mutable double m_value;
   mutable GSLVector m_der;

Code/Mantid/Framework/CurveFitting/src/CostFuncLeastSquares.cpp

@@ -25,7 +25,10 @@ DECLARE_COSTFUNCTION(CostFuncLeastSquares,Least squares)
 /**
  * Constructor
  */
-CostFuncLeastSquares::CostFuncLeastSquares() : CostFuncFitting(),m_value(0),m_pushed(false),
+CostFuncLeastSquares::CostFuncLeastSquares() : CostFuncFitting(),
+  m_includePenalty(true),
+  m_value(0),
+  m_pushed(false),
   m_log(Kernel::Logger::get("CostFuncLeastSquares")) {}
 
 /** Calculate value of cost function
@@ -56,13 +59,16 @@ double CostFuncLeastSquares::val() const
   }
 
   // add penalty
-  for(size_t i=0;i<m_function->nParams();++i)
+  if (m_includePenalty)
   {
-    if ( !m_function->isActive(i) ) continue;
-    API::IConstraint* c = m_function->getConstraint(i);
-    if (c)
+    for(size_t i=0;i<m_function->nParams();++i)
     {
-      m_value += c->check();
+      if ( !m_function->isActive(i) ) continue;
+      API::IConstraint* c = m_function->getConstraint(i);
+      if (c)
+      {
+        m_value += c->check();
+      }
     }
   }
 
@@ -82,9 +88,12 @@ void CostFuncLeastSquares::addVal(API::FunctionDomain_sptr domain, API::Function
 
   double retVal = 0.0;
 
+  double sqrtw = calSqrtW(values);
+
   for (size_t i = 0; i < ny; i++)
   {
-    double val = ( values->getCalculated(i) - values->getFitData(i) ) * values->getFitWeight(i);
+    // double val = ( values->getCalculated(i) - values->getFitData(i) ) * values->getFitWeight(i);
+    double val = ( values->getCalculated(i) - values->getFitData(i) ) * getWeight(values, i, sqrtw);
     retVal += val * val;
   }
   
@@ -273,6 +282,7 @@ void CostFuncLeastSquares::addValDerivHessian(
       std::cerr << std::endl;
     }
   }
+  double sqrtw = calSqrtW(values);
   for(size_t ip = 0; ip < np; ++ip)
   {
     if ( !function->isActive(ip) ) continue;
@@ -281,7 +291,8 @@ void CostFuncLeastSquares::addValDerivHessian(
     {
       double calc = values->getCalculated(i);
       double obs = values->getFitData(i);
-      double w = values->getFitWeight(i);
+      // double w = values->getFitWeight(i);
+      double w =  getWeight(values, i, sqrtw);
       double y = ( calc - obs ) * w;
       d += y * jacobian.get(i,ip) * w;
       if (iActiveP == 0 && evalFunction)
@@ -319,7 +330,8 @@ void CostFuncLeastSquares::addValDerivHessian(
       double d = 0.0;
       for(size_t k = 0; k < ny; ++k) // over fitting data
       {
-        double w = values->getFitWeight(k);
+        // double w = values->getFitWeight(k);
+        double w = getWeight(values, k, sqrtw);
         d += jacobian.get(k,i) * jacobian.get(k,j) * w * w;
       }
       PARALLEL_CRITICAL(hessian_set)
@@ -497,5 +509,23 @@ void CostFuncLeastSquares::calActiveCovarianceMatrix(GSLMatrix& covar, double ep
 
 }
 
+//----------------------------------------------------------------------------------------------
+/** Get weight of data point i(1/sigma)
+  */
+double CostFuncLeastSquares::getWeight(API::FunctionValues_sptr values, size_t i, double sqrtW) const
+{
+  return (values->getFitWeight(i) / sqrtW);
+}
+
+//----------------------------------------------------------------------------------------------
+/** Get square root of normalization weight (W)
+  */
+double CostFuncLeastSquares::calSqrtW(API::FunctionValues_sptr values) const
+{
+  UNUSED_ARG(values);
+
+  return 1.0;
+}
+
 } // namespace CurveFitting
 } // namespace Mantid