SCRO / Journées de l'optimisation

HEC Montréal, 29-31 mai 2023


HEC Montréal, 29 — 31 mai 2023

Horaire Auteurs Mon horaire

OPMI Operations Planning and Management I

29 mai 2023 10h30 – 12h10

Salle: Saine Marketing (vert)

Présidée par Raf Jans

4 présentations

  • 10h30 - 10h55

    Balancing Decision Components in Multi-Objective Optimization

    • Mazyar Zahedi-Seresht, prés., University Canada West
    • Shahrzad Khosravi, University Canada West
    • Bahram Sadeghi Bigham, Alzahra University, Faculty of Mathematical Sciences, Department of Computer Science
    • Erfan Veisi, Continuous Improvement Department, Mirab Valves Company

    Managers often have various indicators to evaluate the performance of units and aim for an optimal state in all of them. However, this rarely happens, and improving one indicator is often accompanied by a decline in another. These events lead each person to evaluate the results according to their own perspective and come up with different answers. Another problem with evaluation is that sometimes some indicators show good performance while others show poor performance, and in short, the results in different indicators are not balanced.
    For example, when a production company reaches a stable and dynamic state, the production and sales process, along with other related components, take a balanced trend. This means that the inventory level is suitable, and we are supposed to prevent the warehouse from becoming too full or empty by maintaining the existing inventory flow. The inventory level should have a balanced and proportional relationship with the company's production. On the other hand, the amount of sales is also linearly and directly related to the inventory and production. In the ideal and balanced state, the number of product packages and the number of exit permits are all directly and linearly related to sales. When the volume of packaging is much less than the volume of sales, we will face an accumulation of goods in the packaging unit and consequently a delay in commitments. If the volume of packaging has very different numbers compared to the volume of permits, this mismatch is a sign of an anomaly and a problem in the organization. Similarly, we can get some information about the ratio of each of these values and compare them in pairs. Managers, to evaluate the indicators and take into account the balance between them, are forced to look at different dimensions at the same time and draw conclusions that in the same situation, each conclusion is different from the next.
    The aim of this article is to present a computational method for indicating the balance of indices. In this method, after standardizing and normalizing the data, all indices are plotted on a radar chart. Then, using a concept in computational geometry called fatness; the fatness value is compared to that of a regular polygon. The resulting value will be less than 1 and can be compared to the fatness of a regular polygon. In this way, a constant number can be calculated to represent the balance for any performance, which has much better results compared to statistical methods.

  • 10h55 - 11h20

    Evaluating the Environmental Benefits of Wood Pellet Supply Chain coordination: A Case Study of Canadian northern communities

    • Zahra Vazifeh, prés., Concordia university
    • Fereshteh Mafakheri, Concordia University
    • chunjiang An, Concordia university

    As the world seeks to reduce greenhouse gas emissions and address the challenge of climate change, bioenergy has become an important part of the energy mix, particularly in the context of decarbonization. Among various types of bioenergy, wood pellets which are produced from wood waste or sawdust, and are utilized for heating and electricity generation has attracted attention globally. The production and consumption of wood pellets have rapidly increased, creating a competitive global market with projections indicating that demand will continue to grow in the foreseeable future. To remain competitive in the market, an efficient management of wood pellet supply chain from harvesting feedstock to bioenergy conversion is very critical. One of the main important tasks of the supply chain management is to coordinate the channel, which is to remove barriers and make sure that every party in the chain work accordingly to optimize objectives of the channel. Coordination of pellet supply chains could play a strategic role in the efficient utilization of existing resources and consequently enhancing the level of energy generation. The present research, with focusing on developing a new contract based coordination mechanism in wood pellet production supply chain, compares the result with both centralized and decentralized decision making structure to indicate the role of coordination in the channel. Utilizing the concept of life cycle assessment in the Stackelberg game, a bi-level nonlinear model with two objective function (environmental and economic) for the leader is suggested, to obtain the equilibrium solution. A case study of wood pellet supply chain, as an alternative option for diesel, in three northern Canadian communities is explored to demonstrate the application of the proposed model. Our findings suggest that coordination can lead to significant emissions reductions, demonstrating the potential for coordinated supply chains to play an important role in achieving sustainability targets.

  • 11h20 - 11h45

    Manufacturing Process Digitization: A Case Study In a SME

    • Angassu Girma Mullisa, Acute industrial consultation & training
    • Walid Abdul-Kader, University of Windsor
    • Walid Abdul-Kader, prés., University of Windsor

    Small and Medium Enterprises (SME) have limited budgets to implement digitization in their manufacturing processes. The widely used value stream mapping (VSM) is a static lean manufacturing tool. A case study is conducted in a SME manufacturing rebar with a capacity of 3021 tons per year. The company sets forth sustainability goals. For example, for total energy consumption, the goal is to reduce it by 20%. For machines, the goal is to reduce the downtime by 30%. For production lead time, the goal is to decrease it by 25%. To improve the performance, this research work considers dynamic/digital value stream mapping (DVSM) by collecting data about four categories of supporting elements: Time elements, quantity elements, energy elements, and maintenance elements. For every supporting element, basic and comprehensive key performance indicators (KPIs) were evaluated and contrasted with the VSM. As initial results, the process lead time dropped by 56%. Other details and results are being analyzed and will be released about the benefit of digitizing the manufacturing process of this rebar company. The cost involved in digitizing the process is negligible compared to the benefits achieved. The findings of this research can have an impact on introducing digitization in SMEs.

  • 11h45 - 12h10

    The lot-sizing problem with simultaneous backlogging and lost sales

    • Gislaine Mara Melega, HEC/CIRRELT Montréal
    • Matthew Arsenault, HEC Montréal
    • Raf Jans, prés., HEC Montréal
    • Julie Paquette, HEC Montréal

    In this study, we extend the classical lot-sizing problem to incorporate simultaneously backlog and lost sales. We assume that if there is a stockout, some proportion of the affected customers will be willing to accept a backlog, while the others will not, resulting in a lost sale. We propose formulations which combine different assumption on backlog and lost sales. We also consider the case with multiple customer types, where customers have a different willingness to wait. To model these problems, we use the facility location reformulation proposed for the lot-sizing problems that provides the flexibility and structure to formulate the problem with both backlog and lost sales decisions simultaneously. We present extensive computational experiments and make comparisons between the different formulations on their performance and structure of the solution. We also propose a relax-and-fix heuristic based on the restricted backlog formulation. We then evaluate the impact that certain parameters have on the performance of the formulations by conducting sensitivity analyses on the following parameters: capacity, customer’s willingness to wait, and lost sales cost.