From Zdravko Cvetkovski's Inequalities — Theorems, Techniques and Selected:
Let $a$, $b$, $c$ and $d$ be positive real numbers. Prove that: $$a^4+b^4+c^4+d^4+2abcd \geq a^2b^2+ a^2c^2+ a^2d^2+ b^2c^2+ b^2d^2+ c^2d^2.$$
Please suggest how to prove this inequality using basic methods. No logarithms.
Assume $a\geq b\geq c\geq d.$ Let $$f(a) = a^4+b^4+c^4+d^4+2abcd - a^2b^2-a^2c^2-a^2d^2-b^2c^2-b^2d^2-c^2d^2.$$ Then, $$f'(a) = 4a^3 -2a(b^2+c^2+d^2)+2bcd$$ and $$f''(a) = 12a^2 - 2b^2-2c^2-2d^2\geq 0.$$ Hence, $f'$ is increasing and $f'(b) = 2(b^3-b(c^2+d^2)+bcd) \geq 2b(b-c)(b-d)\geq 0.$ Thus, $f(a)$ is increasing and therefore:
$$f(a)\geq f(b) = b^4+c^4+d^4+2b^2cd-2b^2c^2-2b^2d^2-c^2d^2 = $$$$b^4-2b^2(c^2+d^2-cd)+c^4+d^4-c^2d^2 = (b^2 - c^2-d^2+cd)^2+2cd(c-d)^2\geq 0.$$
Let $a\geq b\geq c\geq d$.
Thus, $$d(d-a)(d-b)(d+2c)+d(d-a)(d+2b)(d-c)+d(d+2a)(d-b)(d-c)\geq0$$ or $$d^4-(ab+ac+bc)d^2+2abcd\geq0.$$ Thus, it's enough to prove that $$a^4+b^4+c^4\geq(a^2+b^2+c^2-ab-ac-bc)d^2+a^2b^2+a^2c^2+b^2c^2$$ or $$\sum_{cyc}(a^2-b^2)^2\geq\sum_{cyc}(a-b)^2d^2$$ or $$\sum_{cyc}(a-b)^2((a+b)^2-d^2)\geq0,$$ which is obvious.
Here $\sum\limits_{cyc}$ means $\sum\limits_{a\rightarrow b\rightarrow c\rightarrow a}$.
Also, we can use the BW.
Let $a=\min\{a,b,c,d\}$, $b=a+u$, $c=a+v$ and $d=a+w$.
Thus, $u$, $b$ and $w$ are non-negatives and $$\sum_{cyc}a^4+2abcd-\frac{1}{6}\sum_{sym}a^2b^2=$$ $$\sum_{cyc}(3u^2-2uv)a^2+\sum_{cyc}\left(4u^3-u^2v-u^2w+\frac{2}{3}uvw\right)a+\sum_{cyc}(u^4-u^2v^2)\geq0.$$
Also there are proofs by EV Method and by SOS.
To ensure non-negativity, let
$$a =: x_1^2, \qquad b =: x_2^2, \qquad c =: x_3^2, \qquad d =: x_4^2$$
Hence, proving the given inequality is reduced to proving than an octic form is globally non-negative over $\Bbb R^4$. Using Macaulay2 (version 1.16) with package SumsOfSquares:
i1 : needsPackage( "SumsOfSquares" );
i2 : R = QQ[x1,x2,x3,x4];
i3 : f = x1^8 + x2^8 + x3^8 + x4^8 + 2x1^2x2^2x3^2x4^2 - x1^4x2^4 - x1^4x3^4 - x1^4x4^4 - x2^4x3^4 - x2^4x4^4 - x3^4x4^4;
i4 : sosPoly solveSOS f;
Unfortunately, the SOS decomposition produced by Macaulay2 is simply too large to include in this answer, as it exceeds $30000$ characters. I would not call this a basic method. Nonetheless, I find this exercise in mathematical brutality interesting.