Pellet Production Optimization using a Parallelized Progressive Hedging Algorithm

TL;DR

This paper presents a two-stage stochastic programming model for optimizing pellet production under biomass yield and quality uncertainties, utilizing parallelized algorithms to improve computational efficiency.

Contribution

It introduces a novel parallelized progressive hedging algorithm for faster convergence in pellet production optimization under uncertainty.

Findings

Effective optimization of biomass-to-pellet supply chain decisions.

Parallelization schemes significantly reduce computation time.

Model accommodates different international pellet standards.

Abstract

Renewable energy policies have driven the wood pellet market over the last decades worldwide. Among other factors, the return from this business depends largely on how well the producers manage the uncertainty associated with biomass yield and quality. This study develops a two-stage stochastic programming model that optimizes different critical decisions (e.g., harvesting, storage, transportation, quality inspection, and production decisions) of a biomass-to-pellet supply system under biomass yield and quality uncertainty. The goal is to economically produce pellets while accounting for the different pellet standards set forward by the U.S. and European markets. We propose two parallelization schemes to efficiently speed up the convergence of the overall decomposition method. We use Mississippi as a testing ground to visualize and validate the performance of the algorithms.