diff --git a/ParadisEO-GPU/src/problems/eval/QAPIncrEval.h b/ParadisEO-GPU/src/problems/eval/QAPIncrEval.h
new file mode 100644
index 000000000..b9364101a
--- /dev/null
+++ b/ParadisEO-GPU/src/problems/eval/QAPIncrEval.h
@@ -0,0 +1,124 @@
+/*
+ <QAPIncrEval.h>
+ Copyright (C) DOLPHIN Project-Team, INRIA Lille - Nord Europe, 2006-2010
+
+ Karima Boufaras, Thé Van LUONG
+
+ This software is governed by the CeCILL license under French law and
+ abiding by the rules of distribution of free software.  You can  use,
+ modify and/ or redistribute the software under the terms of the CeCILL
+ license as circulated by CEA, CNRS and INRIA at the following URL
+ "http://www.cecill.info".
+
+ As a counterpart to the access to the source code and  rights to copy,
+ modify and redistribute granted by the license, users are provided only
+ with a limited warranty  and the software's author,  the holder of the
+ economic rights,  and the successive licensors  have only  limited liability.
+
+ In this respect, the user's attention is drawn to the risks associated
+ with loading,  using,  modifying and/or developing or reproducing the
+ software by the user in light of its specific status of free software,
+ that may mean  that it is complicated to manipulate,  and  that  also
+ therefore means  that it is reserved for developers  and  experienced
+ professionals having in-depth computer knowledge. Users are therefore
+ encouraged to load and test the software's suitability as regards their
+ requirements in conditions enabling the security of their systems and/or
+ data to be ensured and,  more generally, to use and operate it in the
+ same conditions as regards security.
+ The fact that you are presently reading this means that you have had
+ knowledge of the CeCILL license and that you accept its terms.
+
+ ParadisEO WebSite : http://paradiseo.gforge.inria.fr
+ Contact: paradiseo-help@lists.gforge.inria.fr
+ */
+
+#ifndef __QAPIncrEval_H
+#define __QAPIncrEval_H
+
+#include <eval/moCudaEvalFunc.h>
+
+/**
+ * Incremental Evaluation of QAP
+ */
+
+template<class Neighbor>
+class QAPIncrEval: public moCudaEvalFunc<Neighbor> {
+
+public:
+
+	typedef typename Neighbor::EOT EOT;
+	typedef typename EOT::Fitness Fitness;
+
+	/**
+	 * Constructor
+	 */
+
+	QAPIncrEval() {
+	}
+
+	/**
+	 * Destructor
+	 */
+
+	~QAPIncrEval() {
+	}
+
+	/**
+	 * functor of  incremental evaluation of the solution(function inline can be called from host or device)
+	 * @param _sol the solution to evaluate
+	 * @param _fitness the fitness of the current solution
+	 * @param _index the set of information to compute fitness neighbor
+	 * @param _index[0] the first position of swap
+	 * @param _index[1] the second position of swap
+	 * @param _index[2] thesolution size
+	 * @param _index[3] the id of neighbor
+	 */
+
+inline __host__ __device__ Fitness operator() (EOT & _sol,Fitness _fitness, unsigned int *_index) {
+
+	Fitness tmp;
+
+	/* dev_a & dev_b are global device variable, data specific to QAP problem (flow & distance matices)
+	 * _index[0] the first position of swap
+	 * _index[1] the second position of swap
+	 * _index[2] thesolution size
+	 * _index[3] the id of neighbor
+	 */
+
+	tmp=_fitness+compute_delta(dev_a,dev_b,_sol,_index[0],_index[1],_index[2],_index[3]);
+	return tmp;
+
+}
+
+    /**
+	 * compute the new fitness of the solution after permutation of pair(i,j)(function inline called from host  device)
+	 * @param _a the flow matrix of size*size (specific data of QAP problem must be declared as global device variable)
+	 * @param _b the distance matrix of size*size (specific data of QAP problem must be declared as global device variable)
+	 * @param _sol the solution to evaluate
+	 * @param _i the first position of swap
+	 * @param _j the second position of swap
+	 * @param _size the id of neighbor
+	 * @param _id the neighbor identifier
+	 */
+
+inline __host__ __device__  int compute_delta(int * _a,int * _b,EOT & _sol, int _i, int _j,int _size, int id){
+
+  int d;
+  int k;
+
+  d = (_a[_i*_size+_i]-_a[_j*_size+_j])*(_b[_sol[_j+id*_size]*_size+_sol[_j+id*_size]]-_b[_sol[_i+id*_size]*_size+_sol[_i+id*_size]]) +
+    (_a[_i*_size+_j]-_a[_j*_size+_i])*(_b[_sol[_j+id*_size]*_size+_sol[_i+id*_size]]-_b[_sol[_i+id*_size]*_size+_sol[_j+id*_size]]);
+
+
+  for (k = 0; k < _size; k=k+1)
+    if (k!=_i && k!=_j)
+
+      d = d + (_a[k*_size+_i]-_a[k*_size+_j])*(_b[_sol[k+id*_size]*_size+_sol[_j+id*_size]]-_b[_sol[k+id*_size]*_size+_sol[_i+id*_size]]) +
+      (_a[_i*_size+k]-_a[_j*_size+k])*(_b[_sol[_j+id*_size]*_size+_sol[k+id*_size]]-_b[_sol[_i+id*_size]*_size+_sol[k+id*_size]]);
+
+  return(d);
+}
+
+};
+
+#endif