TORS'23

Supply chain design (SCD) problems deal with strategic decisions on the reengineering of the network structure, in terms of deployment of long-term resources and the planning of the role of those resources' and their mission in terms of capacities to assign and policy-making for the primary supply chain (SC) activities. Such strategic investments in resources, policy making and deployment issues are the responsibility of top management and at a core of corporate strategy.

From an operations research perspective, the design of SC networks is one of the most important areas of application for multi-commodity network design models. Despite the abundant literature in this area, extant approaches fail to comprehensively capture the reality of modern SC systems.

In this paper, based on our review of literature in the field, unifying SCD and SC network design models, we identify critical challenges and main opportunities for future research in this area. These are articulated around three axes, with the aim to improve the ability of SCD models to drive business decisions: (1) expanding of the scope of SCD models, (2) enhancing methodologically the accuracy of SCD models and (3) improving solution methods and decision-making approaches. In each of these axes, we review the current state of the art of the literature, and we confront this to the requirements of the modern supply chain systems. We identify main gaps in the extant literature, provide guidelines on key modeling-solvability tradeoffs and identify concrete opportunities for future research fostering the robustness of SC designs.

We consider an one machine scheduling model, minimizing a classical objective function – either the total completion time or the maximum tardiness – and with two sets of jobs : one with initial jobs already scheduled and one with new jobs which must be inserted in the schedule. As such reschedule can create a modification of the schedule of the initial jobs, a disruption objective is considered in addition to the original objective. This additional objective can be formulated in four different ways. Such model has been introduced by Hall and Potts, minimizing either a linear aggregation of the two objectives or the initial objective under a constraint giving an upper limit of the disruption objective. In this study, the aim is to obtain the set of efficient schedules in regard of the two objectives. Algorithms are provided for the eight possible bi-objective problems and illustrated by some didactic examples.

In this talk, we will discuss critical challenges in sustainable and responsive city logistics and last mile delivery. We will cover research progress on the adoption of emerging transportation technology such as drones, electrical vehicles, and crowdsourcing, and on novel data-driven optimization approaches to address the uncertainty prevalent due to technology and operating environments.

Sustainable decision-making problems are complex in a context where decisions have to be made based concurrently on economic, social and environmental considerations and to be evaluated on the short, medium and long term. In the last five years, development of temporal MCDA has witnessed the interest of some researchers. However, the latest developments remain limited and only a few research studies offer aggregation frameworks for temporal evaluations settings. This tutorial presents recent temporal MCDA methods that we proposed to recommend sustainable decisions. Three MCDA methods will be presented. First, the outranking MUPOM method will demonstrate how outranking MCDA methods, based on concordance-discordance principles, can be generalized to processing temporal impacts of decisions. Secondly, the PROMETHEE-MP method which consists of a multi-period generalization of PROMETHEE under random uncertainty is discussed. Thirdly, we present a temporal sorting MCDA method under stochastic uncertainty, called TSMAA-Tri. This proposed method consists of sorting the alternatives into predefined categories based on a generalization of SMAA-Tri to a temporal context (multi-period evaluation of alternatives) where alternative evaluations are stochastic. The proposed methods are illustrated for sustainable forest management to show their applicability.

Keywords: Multiple criteria decision aid (MCDA); Temporal evaluations; Outranking methods; Sorting methods, Sustainable decisions; Uncertainty.

Call centers are complex service systems with labor-intensive operations. They employ millions of agents across the globe and serve as a primary customer-facing channel for firms in many different industries. New technology-driven innovations in call centers are multiplying the opportunities to make more efficient use of agents as they can handle different types of workflow, including inbound calls, outbound calls, emails, web calls, chats and social medias. However, several issues on the management of call center operations emerged also as a result of advanced technology. Managers in practice are particularly interested in how to efficiently share the agent time between the available job types in order to improve the call center performance, i.e., efficient routing mechanisms of multiple flows from multiple channels to the capacity.

We discuss various optimization problems for modern call centers with human-human channels. We illustrate how stochastic models, such as queueing and Markov chain techniques, can help in generating solutions and insights for the settings of delay announcement, chat online assistance, and advertising during waiting. This contributes to bridge the gap between the scientific literature on stochastic modeling and the needs of call center managers in real-life.

In production systems, hierarchical interdependencies of decisions, e.g. between tactical (resource-level) and operational (activity-level) decisions, arise and can be exploited to provide additional gain when optimization is able to integrate these decisions into a combined model. Considering more complex business environments, such as airlines and hospitals, planning problems and their interdependencies are more intricate, since humans are integrated at customer service (demand) side, such as patients in hospitals, or at resource (supply) side, such as crew personal in airlines. The planning of their activities (treatments/flights) should comply with working rules in case of crew (duties and pairings) and with medical/clinical standards in case of patients (patient pathways). To this side of complexity related to composition of activity structures, one faces another complexity of multiplicity of these structures within (patient/crew) flow management/problems.

We present a new two-dimensional aggregation scheme w.r.t designed crossed dimensions for such complexities. This scheme serves as modeling, analytics and optimization design tool for complex systems’ businesses. Assigning appropriate aggregation/granulation levels to each complexity dimension, one can easily characterize different decision problems at strategic, tactical and operational levels – some of those may have not been apparent in advance. Types of solution methods can also be entailed in a generic way: Most problems at aggregated (low-level detail) levels require descriptive/predictive analytics as well as AI for learning complex structures. However, planning problems at granulated (high-level detail) levels should be tackled by prescriptive analytics and OR. Further, interdependencies between aggregated and granulated levels of decision problems give rise to AI/OR synergies of methodologies.