mutSubgraphMST {mcMST}

Subgraph-mutator for edge list representation.

Description

mutSubgraphMST selects a random edge e = (u, v) and traverses the tree starting form u and v respectively until a connected subtree of at most sigma edges is selected. Then the subtree is replaced with the optimal spanning subtree regarding one of the objectives with equal probability.

Usage

mutSubgraphMST(
  ind,
  sigma = floor(ncol(ind)/2),
  scalarize = FALSE,
  instance = NULL
)

Arguments

ind

[matrix(2, m)]
Matrix of edges (each column is one edge).

sigma

[integer()]
Upper bound for the size of the selected subtree.

scalarize

[logical(1)]
Should a scalarized version of the the subproblem be solved? If TRUE, a random weight \lambda \in [0,1] is sampled from a U[0, 1]-distribution. Next, a weighted sum scalarization \lambda \cdot c_1 + (1 - \lambda) \cdot c_2 of the subproblem is solved. Default is FALSE, i.e., the single-objective subproblem is solved. The objective to optimize for is sampled with equal probability.

instance

[grapherator]
Multi-objective graph.

Value

[matrix(2, m)] Mutated edge list.

See Also

Evolutionary multi-objective algorithm mcMSTEmoaBG

Other mcMST EMOA mutators: mutEdgeExchange(), mutKEdgeExchange(), mutSubforestMST(), mutUniformPruefer()