Scheduling tasks in environments with unobserved state changes (e.g., machine failures, variable job processing times, human fatigue) is challenging. This paper presents a Hidden Markov Model (HMM) framework to represent latent operational states that influence schedule performance. We show how Viterbi inference estimates the most probable current state, enabling lookahead schedule adjustments. Simulations on a single-machine scheduling problem with hidden deterioration states demonstrate reduced tardiness compared to non-adaptive baselines.
When viewing an HMM schedule, several specific terms and columns dictate the logistics plan: hmm schedule