LibraryReExportIsNPComplete

From APIDesign

Revision as of 09:47, 11 February 2009

This page describes a way to convert any wikipedia::3SAT problem to a solution of finding the right configuration from conflicting libraries in a system that can re-export APIs. Thus proving that the later problem is wikipedia::NP-complete.

wikipedia::3SAT

The problem of satisfying a logic formula remains NP-complete even if all expressions are written in wikipedia::conjunctive normal form with 3 variables per clause (3-CNF), yielding the 3SAT problem. This means the expression has the form:

Failed to parse (unknown error): (x_{11} \vee x_{12} \vee x_{13}) \wedge

Failed to parse (unknown error): (x_{21} \vee x_{22} \vee x_{23}) \wedge

Failed to parse (unknown error): (x_{31} \vee x_{32} \vee x_{33}) \wedge

Failed to parse (unknown error): ...

Failed to parse (unknown error): (x_{n1} \vee x_{n2} \vee x_{n3})

where each Failed to parse (unknown error): x_{ab}

is a variable Failed to parse (unknown error): v_i
or a negation of a variable Failed to parse (unknown error): \neg v_i

. Each variable Failed to parse (unknown error): v_i

can appear multiple times in the expression.

Module Dependencies Problem

Let Failed to parse (unknown error): A, B, C, ...

denote an API.

Let Failed to parse (unknown error): A_1, A_{1.1}, A_{1.7}, A_{1.11}

denote compatible versions of API Failed to parse (unknown error): A

.

Let Failed to parse (unknown error): A_1, A_{2.0}, A_{3.1}

denote incompatible versions of API Failed to parse (unknown error): A

.

Let Failed to parse (unknown error): A_{x.y} > B_{u.v}

denote the fact that version x.y of API A depends on version u.v of API B.

Let Failed to parse (unknown error): A_{x.y} \gg B_{u.v}

denote the fact that version x.y of API A depends on version u.v of API B and that it re-exports B's elements.

Let Repository Failed to parse (unknown error): R=(M,D)

be any set of modules with their various versions and their dependencies on other modules with or without re-export.

Let C be a Configuration in a repository Failed to parse (unknown error): R=(M,D) , if Failed to parse (unknown error): C \subseteq M , where following is satisfied:

Failed to parse (unknown error): \forall A_{x.y} \in C, \forall A_{x.y} \gg B_{u.v} \in D \Rightarrow \exists w >= v \wedge B_{u.w} \in C

- each re-exported dependency is satisfied with some compatible version

Failed to parse (unknown error): \forall A_{x.y} \in C, \forall A_{x.y} > B_{u.v} \in D \Rightarrow \exists w >= v B_{u.w} \in C

- each dependency is satisfied with some compatible version

Let there be two chains of re-exported dependencies Failed to parse (unknown error): A_{p.q} \gg ... \gg B_{x.y}

and Failed to parse (unknown error): A_{p.q} \gg ... \gg B_{u.v}
then Failed to parse (unknown error): x = u \wedge y = v
- this guarantees that each class has just one, exact meaning for each importer

Module Dependency Problem

Let there be a repository Failed to parse (unknown error): R=(M,D)

and a module Failed to parse (unknown error): A \in M

. Does there exist a configuration Failed to parse (unknown error): C

in the repository Failed to parse (unknown error): R

, such that the module Failed to parse (unknown error): A \in C , e.g. the module can be enabled?

Converstion of wikipedia::3SAT to Module Dependencies Problem

Let there be wikipedia::3SAT formula with with variables Failed to parse (unknown error): v_1, ..., v_m

as defined above.

Let's create a repository of modules Failed to parse (unknown error): R . For each variable Failed to parse (unknown error): v_i

let's create two modules Failed to parse (unknown error): M^i_{1.0}
and Failed to parse (unknown error): M^i_{2.0}

, which are mutually incompatible and put them into repository Failed to parse (unknown error): R .

For each formula Failed to parse (unknown error): (x_{i1} \vee x_{i2} \vee x_{i3})

let's create a module Failed to parse (unknown error): F^i

that will have three compatible versions. Each of them will depend on one variable's module. In case the variable is used with negation, it will depend on version 2.0, otherwise on version 1.0. So for the formula 

Failed to parse (unknown error): v_a \vee \neg v_b \vee \neg v_c

we will get:

Failed to parse (unknown error): F^i_{1.1} \gg M^a_{1.0}

Failed to parse (unknown error): F^i_{1.2} \gg M^b_{2.0}

Failed to parse (unknown error): F^i_{1.3} \gg M^c_{2.0}

All these modules and dependencies add into repository Failed to parse (unknown error): R

Now we will create a module Failed to parse (unknown error): T_{1.0}

that depends all formulas: 

Failed to parse (unknown error): T_{1.0} \gg F^1_{1.0}

Failed to parse (unknown error): T_{1.0} \gg F^2_{1.0}

...

Failed to parse (unknown error): T_{1.0} \gg F^n_{1.0}

and add this module as well as its dependencies into repository Failed to parse (unknown error): R .

Claim: There Failed to parse (unknown error): \exists

a configuration Failed to parse (unknown error): C
of repository Failed to parse (unknown error): R
and Failed to parse (unknown error): T_{1.0} \in C
Failed to parse (unknown error): \Longleftrightarrow
there is a solution to the wikipedia::3SAT formula.

Proof

"Failed to parse (unknown error): \Leftarrow ": Let's have an evaluation of each variable to either true or false that evaluates the whole wikipedia::3SAT formula to true. Then

Failed to parse (unknown error): C = \{ T_{1.0} \} \bigcup

Failed to parse (unknown error): \{ M^i_{1.0} : v_i \} \bigcup \{M^i_{2.0} : \neg v_i \} \bigcup

Failed to parse (unknown error): \{ F^i_{1.1} : x_{i1} \} \bigcup \{ F^i_{1.2} : \neg x_{i1} \wedge x_{i2} \} \bigcup \{ F^i_{1.3} : \neg x_{i1} \wedge \neg x_{i2} \wedge x_{i3} \}

It is clear from the definition that each Failed to parse (unknown error): M^i

and Failed to parse (unknown error): F^i
can be in the Failed to parse (unknown error): C
just in one version. Now it is important to ensure that each module is present always at least in one version. This is easy for Failed to parse (unknown error): M^i
as its Failed to parse (unknown error): v_i
needs to be true or false, and that means one of Failed to parse (unknown error): M^i_{1.0}
or  Failed to parse (unknown error): M^i_{2.0}
will be included. Can there be a Failed to parse (unknown error): F^i
which is not included? Only if Failed to parse (unknown error): \neg x_{i1} \wedge \neg x_{i2} \wedge \neg x_{i3}
but that would mean the whole or would evaluate to false and as a result also the wikipedia::3SAT formula would evaluate to false. This means that dependencies of Failed to parse (unknown error): T_{1.0}
on Failed to parse (unknown error): F^i
modules are satisfied. Are also dependencies of every Failed to parse (unknown error): F^i_{1.q}
satisfied? From all the three versions, there is just one Failed to parse (unknown error): F^i_{1.q}

, the one its Failed to parse (unknown error): x_{iq}

evaluates to true. However Failed to parse (unknown error): x_{iq}
can either be without negation, and as such Failed to parse (unknown error): F^i_{1.q}
depends on Failed to parse (unknown error): M^j_{1.0}
which is included as Failed to parse (unknown error): v_j
is true. Or Failed to parse (unknown error): x_{iq}
contains negation, and as such Failed to parse (unknown error): F^i_{1.q}
depends on Failed to parse (unknown error): M^j_{2.0}
which is included as Failed to parse (unknown error): v_j
is false. qed.