An Expectation-Maximization Algorithm-based Autoregressive Model for the Fuzzy Job Shop Scheduling Problem

• DOI: 10.48550/arxiv.2502.00018
• Publication Date: 2025-01-11
• Authors: Yijian Wang; T.-Y. Guo; Zhaoqiang Liu
• Keywords: Maximization

The fuzzy job shop scheduling problem (FJSSP) emerges as an innovative extension to the (JSSP), incorporating a layer of uncertainty that aligns more closely with complexities real-world manufacturing environments. This improvement increases computational complexity deriving solution while improving its applicability. In domain deterministic scheduling, neural combinatorial optimization (NCO) has recently demonstrated remarkable efficacy. However, application realm been relatively unexplored. paper aims bridge this gap by investigating feasibility employing networks assimilate and process information for resolution FJSSP, thereby leveraging advancements in NCO enhance methodologies. To achieve this, we approach FJSSP generative task introduce expectation-maximization algorithm-based autoregressive model (EMARM) address it. During training, our alternates between generating schemes from given instances (E-step) adjusting weights based on these generated (M-step). novel methodology effectively navigates around substantial hurdle obtaining ground-truth labels, which is prevalent issue frameworks. testing, experimental results demonstrate superior capability EMARM addressing showcasing effectiveness potential practical applications scheduling.