Let the incircle of $\triangle ABC$ meet the sides at points $A'$, $B'$, $C'$ (with $A$ opposite $A'$, etc). Tangent segments being congruent, we can define lengths $a$, $b$, $c$ where
$$a := |AB'|=|AC'| \qquad b := |BC'| = |BA'| \qquad c := |CA'| = |CB'|$$

Swapping the sub-segments on each side determines points $A''$, $B''$, $C''$ such that paths along the boundary from $A$ to $A''$, from $B$ to $B''$, and from $C$ to $C''$ each has length $a+b+c$: the new points are the desired perimeter-bisecting points.
Invoking Ceva's Theorem, we find
$$\frac{|A''C|}{|BA''|} \cdot \frac{|B''A|}{|CB''|} \cdot \frac{|C''B|}{|AC''|} = \frac{c}{b} \cdot \frac{a}{c} \cdot \frac{b}{a} = 1$$
guaranteeing that cevians $AA''$, $BB''$, $CC''$ meet at a point. $\square$
Note. The argument is effectively the same as that of @RandomMathEnthusiast. RME's values $\frac12(-a+b+c)$, $\frac12(a-b+c)$, $\frac12(a+b-c)$ are precisely the lengths of the incircle-tangent segments when the sides of the triangle are $a$, $b$, $c$, but obtained algebraically.
Note 2. YNK notes in a comment to the question that $A''$, $B''$, $C''$ are points of tangency with the triangle's excircles. This old answer of mine provides a justification of this fact.
Note 3. YNK indicates in a comment this answer that the point of concurrence is the Nagel point, indexed as $X_8$ in the Encyclopedia of Triangle Centers.
As it happens, cevians $AA'$, $BB'$, $CC'$ also concur, at the Gergonne point, aka $X_7$ in the ETC.
The fact that $A'$ and $A''$ (likewise, $B'$ and $B''$, and $C'$ and $C''$) are mutual reflections in the midpoint of the side of the triangle makes $X_7$ and $X_8$ "isotomic conjugates".