The stochastic bilevel continuous knapsack problem with uncertain follower's objective