I have doubt in computing IGD value in multi-objective optimization. In IGD, Euclidean distance is computed. My doubt is that Euclidean distance is computed between what? whether compute Euclidean distance between Pareto optimal solutions(denoted by PS*={x| x is a Pareto optimal solution}) and obtained set (ours) or between Pareto front (denoted PF*={(f1(x),f2(x),....,fM(x))|x is Pareto optimum}) and our obtained Pareto front PF.