Weighted Costs

solve_lap_weighted multiplies each cost entry by a per-element weight before solving — useful when you want to bias the assignment (e.g. down-weighting a noisy feature-matching cost, or emphasizing a particular sub-region of the matrix) without discarding the original cost values.

import numpy as np
import fastlap

cost    = np.array([[1, 2], [3, 4]], dtype=np.float64)
weights = np.array([[1, 0.5], [0.5, 1]], dtype=np.float64)

total, rows, cols = fastlap.solve_lap_weighted(cost, weights, algorithm="lapjv")

The returned cost is unweighted

The solver optimizes over weight[i][j] * cost[i][j], but total_cost is always computed from the original, unweighted cost_matrix. This matters: the assignment reflects your weighting preference, but the reported cost stays meaningful in the original units (e.g. real distance, real dollars).

cost_matrix and weights must have the same shape — a mismatch raises ValueError.

Parameters

Same shape as solve_lap, plus the weights argument:

fastlap.solve_lap_weighted(
    cost_matrix,
    weights,
    algorithm="lapjv",
    maximize=False,
    cost_limit=None,
)

cost_limit gating (see Cost Limit) is applied against the original cost values, consistent with the returned total_cost:

costs = np.array([[2.0, 50.0], [50.0, 20.0]])
weights = np.ones_like(costs)
cost, rows, cols = fastlap.solve_lap_weighted(costs, weights, algorithm="lapjv", cost_limit=10.0)
print(rows)  # [0, None]
print(cost)  # 2.0