- #1
JohnDoe2013
- 5
- 0
One reduction from the 3-SAT problem to 3-COLOR uses an OR-gadget that is basically a set of nodes connected to have the properties of an OR-Gate.
The gadget is 3-colorable regardless whenever all 3 input literals have the same color (regardless of whether the colors mean True or False).
The output node is always forced to be the same color as the color of the literals.
Since, we want the gadget to be 3-colorable only when the color means True, we connect the output node to a "constraint" node that is colored "False".
However, this means that the input graph to the 3-COLOR problem is allowed to have some nodes already colored.
In the definition of thec 3-COLOR problems I have seen, there is no mention that this is allowed. Of course, it does not say that it is not allowed.
Is there some agreement on whether it is allowed?
The gadget is 3-colorable regardless whenever all 3 input literals have the same color (regardless of whether the colors mean True or False).
The output node is always forced to be the same color as the color of the literals.
Since, we want the gadget to be 3-colorable only when the color means True, we connect the output node to a "constraint" node that is colored "False".
However, this means that the input graph to the 3-COLOR problem is allowed to have some nodes already colored.
In the definition of thec 3-COLOR problems I have seen, there is no mention that this is allowed. Of course, it does not say that it is not allowed.
Is there some agreement on whether it is allowed?