Backtracking to get 2 sums as similar as possible

It happens that I have to deliver a code which I must solve with backtracking, but I have gone hand in hand the way of adding and differentiating the sums to compare them. The problem is that; Given a list of numbers, of size X, you have to make 2 lists, each list must be of size X / 2 and at the same time each one must add the closest possible quantity to each other. axis;

lista={1,3,5,7};

n=4; //(length de lista).

the result will be:

lista1={1,7};  y lista2={3,5};

the sum of the numbers in list1 = 8, as well as list2. That sum of the 2 resulting lists should be as similar as possible. advance but it is not very good to unite all the ideas.

public static int[] lista;
public static int n;

public static int suma1;
public static int suma2;////


public static boolean esPosible(int pos1, int pos2) {
    if (pos1 >= 0 && pos2 >= 0 && pos1 < n && pos2 < n) {
        return true;
    }
    return false;
}

public static void prueba(int suma1, int suma2, int pos1, int pos2) {
    StdOut.println(n);
    int mitad1 = 0, mitad2 = 0;
    if (n % 2 == 0) {
        mitad2 = mitad1 = n / 2;
    } else {
        mitad1 = (n + 1) / 2;
        mitad2 = (n - 1) / 2;
    }

    if ((suma1 - suma2) == 0) {
        StdOut.println("el valor de las sumas para el caso sera; suma 1= " + suma1 + " y suma 2= " + suma2);
        return;
    } else {
        if (esPosible(pos1, pos2)) {

            if (pos1 == n) {
                pos1 = 0;
                prueba(suma1, suma2, pos1 + 1, pos2 + 1);
            }


            suma1 = lista[pos1] + lista[pos2];

            prueba(suma1, suma2, pos1 + 1, pos2);

        }
    }

help!