TA4 - Manufacturing

The Integrated Lot Sizing and Cutting Stock Problem with Multiple Modes consists of cutting one-dimensional objects of varying thickness into items for manufacturing products using multiple production modes over a multi-period horizon. It aims to minimize total raw material and inventory holding costs through mathematical modeling and heuristics.

A product's design implicitly defines a set of precedence constraints that govern the ordering of its assembly tasks. These constraints can prevent low-cost manufacturing configurations from being reached by limiting the number of feasible assembly sequences. In the case of assembly line balancing, this issue arises in two distinct contexts. The first is in the validation of an initial set of precedence constraints of a new product that were first generated automatically from methods such as CAD analysis or assembly by disassembly. The second context is the evaluation of an existing precedence graph by designers, to find opportunities to improve the line balance by modifying or removing precedence relations. To address both contexts, we propose a decision aid for the Type-1 Simple Assembly Line Balancing Problem. We develop both heuristic and machine learning-driven methods that are able to quickly predict the impact of a precedence constraint on the line balance. Numerical experiments show that our methods outperform the state-of-the-art methods in precedence constraint validation.

Precast lattice joist production is order-driven, requiring weekly planning of quantities, lengths, and deadlines. This paper studies an integrated lot-sizing and cutting stock problem with purchasing and inventory decisions, proposing two mathematical formulations and heuristics to minimize waste, setups, and inventory costs, validated through computational experiments

We evaluate the impact of customer-driven demand substitution in lot-sizing problems. In each period, customers successively visit a retail location and make purchases based on their preferences. We model the customers' preferences using choice models. The objective is to determine a production plan that maximizes sales profit. Numerical experiments show that our model returns more profitable production plans compared to standard lot-sizing models.