02:30 PM - 02:52 PM
A Comparative Study of Simulation Models for the Production Control of Unreliable Manufacturing Systems
This paper deals with the production control policy of an unreliable manufacturing system producing one part type, and subject to random failures and repairs. The applied production control policy is based on the so-called hedging point policy (HPP), which consists in building and maintaining a safety stock of product in order to fulfill the demand, and protect the production system against shortages during maintenance actions. The main objective of the study is to determine the most efficient option of the ARENA simulation software that simulates properly the production systems under consideration. To this end, four simulation models mimicking the dynamics and the stochastic behavior of the proposed manufacturing system were developed. Concepts of discrete and continuous simulation and modules from the ARENA flow process template, are applied to develop the models. The hedging point policy is used as input parameter of the simulation models, where we seek to determine the optimal production threshold that minimizes the inventory and backlog cost. Based on simulation results, the performance of the models is evaluated, in terms of accuracy and time economy. The obtained results shows that the continuous simulation model that uses C++ inserts outperforms the other models.
02:52 PM - 03:14 PM
Improving Sawmill Agility through Log Classification
Sawmill production is characterized by divergent processes and coproduction. In this context, it is difficult for a production manager to establish a production plan which meets customer demand. This is one of the reasons why the North American lumber industry produces with a mainly make-to-stock strategy. The aim of this research is to evaluate the impact that a better classification of the raw material (logs) would have on sawmill agility. Using a mathematical model to create production plans, we evaluate the performance of the mill to meet the demand in light of the knowledge it has of the raw material.
03:14 PM - 03:36 PM
Combining Stochastic Modeling and Linear Progamming to Improve Aggregate Production Planning
Mathematical models for Aggregate Production Planning (APP) typically omit the dynamics of the underlying production system due to variable workload levels since they assume fixed capacity buffers and predetermined lead times. Pertinent approaches to overcome these drawbacks are either restrictive in their modeling capabilities or prohibitive in their computational effort. In this paper, we introduce an Aggregate Stochastic Queuing (ASQ) model to anticipate capacity buffers and lead time coefficients for the APP model. The ASQ model allows for flexible modeling of the underlying production system and the corresponding optimization algorithm is computationally very well tractable. The APP and the ASQ model are integrated into a hierarchical framework and are solved iteratively. A numerical example is used to highlight the benefits of this novel approach.
03:36 PM - 03:58 PM
Stochastic Scheduling with Abandonments
In this paper, we address the problem of dynamically scheduling jobs with abandonments. Processing times and release dates are arbitrarily distributed while patience times are exponentially distributed. The objective is to minimize holding costs and abandonment costs, either in the class of static list scheduling policies or in the class of dynamic policies with preemption. We first show an equivalence between holding costs and abandonment costs. When processing times are exponentially distributed and all jobs are available at time 0, we provide conditions under which a strict priority rule is optimal in the class of static list scheduling policies. Then we extend this result to the class of dynamic scheduling policy with preemption and arbitrary stochastic release dates.