“…Here • is the Hadamard product, T = (T ij ) = (1∕( i + d j )), and u and v satisfy the vector equations In recent years, efficient numerical algorithms for finding the minimal positive solution of Equation 1 have been extensively studied (e.g., see previous studies 2,[4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] ). On the other hand, these algorithms rely on floating point arithmetic, so that rounding errors are included in the obtained result.…”