PLS (complexity)

In computational complexity theory, PLS, which stands for Polynomial Local Search, is a syntactic subclass of TFNP. An instance of PLS can be given in terms of polynomial algorithms that determine an exponentially sized graph. Given an input x and a vertex in the graph, the polynomial algorithms can compute the neighbors of the vertex in the graph and the cost of the vertex in the graph.

An algorithm for a PLS problem, given an input x, will find a vertex in the graph such that no neighbor has better cost. This class is a subclass of TFNP, because every dag (Directed Acyclic Graph) has a sink.

Examples of PLS-complete problems include local-optimum relatives of TSP, MAXCUT and SAT.

References

• .

Read More