Approximate scenario solutions in the progressive hedging algorithm. A numerical study with an application to fisheries management (Q1176857)

This paper focuses on a time-discrete controllable process in time stages \(t=0,1,\dots,T\). The state of the process at time \(t\) is denoted by the variable \(x_ t\). The transition from the state at time \(t\) to that at time \(t+1\) is governed by a control variable \(u_ t\) but is also dependent on an auxiliary variable, the scenario \(s^ t\). The scenarios can be illustrated with a tree structure, the scenario tree. Thus we can describe the transition as follows: \(x_{t+1}=G_ t(x_ t,u_ t,s^ t)\). Probabilities are attached to the scenarios. To solve a stochastic optimal control problem the so-called `progressive hedging algorithm' developed by \textit{R. T. Rockafellar} and \textit{R. J.-B. Wets} [Math. Oper. Res. 16, No. 1, 119-147 (1991; Zbl 0729.90067)] is specialized. The paper describes how the scenario aggregation principle can be combined with approximate solutions of the individual scenario problems, resulting in a computationally efficient algorithm where two individual Lagrangian-based procedures are merged into one. Computational results are given for an example from fisheries management. Numerical experiments indicate that only crude scenario solutions are needed.