scispace - formally typeset
Search or ask a question
Book ChapterDOI

An Online Decision Support Framework for Managing Abnormal Supply Chain Events

TL;DR: A new agent-based online decision support framework for disruption management, which includes monitoring the KPIs, detecting the root cause for the deviation of KPIS, identifying rectification strategies, finding the optimal rectification strategy and rescheduling operates as necessary in response to the disruption.
Abstract: Enterprises today have acknowledged the importance of supply chain management to achieve operational efficiency, and cutting costs while maintaining quality. Uncertainties in supply, demand, transportation, market conditions, and many other factors can interrupt supply chain operations, causing significant adverse effects. These uncertainties motivate the development of simulation models and decision support system for managing disruptions in the supply chain. In this paper, we propose a new agent-based online decision support framework for disruption management. The steps for disruption management are: monitoring the KPIs, detecting the root cause for the deviation of KPIs, identifying rectification strategies, finding the optimal rectification strategy and rescheduling operates as necessary in response to the disruption. The above framework has been implemented as a decision support system and tested on a refinery case study.

Content maybe subject to copyright    Report

Citations
More filters
Journal ArticleDOI
TL;DR: A comprehensive review of supply chain literature is provided that addresses problems relevant to SCRM using approaches that fall within the AI spectrum and proposes directions for future research at the confluence of SCRM and AI.
Abstract: Supply chain risk management (SCRM) encompasses a wide variety of strategies aiming to identify, assess, mitigate and monitor unexpected events or conditions which might have an impact, mostly adve...

355 citations


Cites background from "An Online Decision Support Framewor..."

  • ...Bansal et al. (2005) propose a similar agent-based approach that monitors key performance indicators (KPIs) to identify disruptions, then evaluates corrective actions and finds the optimal one....

    [...]

  • ...Bansal et al. (2005) propose a similar agent-based approach that monitors key performance indicators (KPIs) to identify disruptions, then evaluates corrective actions and finds the optimal one. Note that these works are the first approaches in this survey so far to offer a holistic approach. Multi-agent systems have also been used as simulation tools for inventory management to reduce costs and improve fill rate (Chan and Chan 2006), to resolve collaboration issues among supply chain entities that arise due to uncertain supply and demand (Kwon, Im, and Lee 2007) and to reduce costs and the bullwhip effect in a multi-stage supply chain (Zarandi, Pourakbar, and Turksen 2008). In all the aforementioned approaches, agent technology is preferred over mathematical optimisation due to its inherent capability to capture negotiation and coordination among different parties and the relatively less computational effort required. In contrast, Mele et al. (2007) propose a hybrid approach for retrofitting a production/distribution supply chain design to address uncertain demand, transport and processing times. A genetic algorithm is employed to solve a multi-stage stochastic model of the problem, but the fitness of each iteration’s individuals is calculated using agent-based simulation. The work of Giannakis and Louis (2011) is notable as it goes beyond simulation to explore the learning capabilities of multi-agent systems in the context of SCRM. The proposed framework addresses all SCRM phases and starts similarly to others by monitoring KPIs to identify deviations, which are then assessed in order to pinpoint the root cause. Through case-based learning, successful past decisions are leveraged when similarities with the current situation are identified (e.g. with regard to root causes). The optimal decision is selected and its results simulated and quantified, with all the information stored in order to be considered in similar situations in the future. Recent work by Blos, da Silva, and Wee (2018) proposes a holistic disruption management framework that combines multi-agent systems with variants of Petri Nets called Production Flow Systems and Place-Transition Petri Nets. The latter are used to model complex supply chain networks in several disruption scenarios. An agent-based system is used to generate more scenarios based on the existing ones, choose the most appropriate mitigation actions and evaluate the resulting effect in performance. Then, based on a risk event database, the supply chain is monitored in order to detect disruptions and immediately propose mitigation plans. As shown in Table 3, the aforementioned agent-based approaches usually focus on very few parameters (up to 4), since it is difficult and computationally complex to simultaneously support a large number of policies that formalise agent behaviour. Similarly to previous categories, large datasets and predictive or learning capabilities are not supported, with the exceptions of the learning capability of Giannakis and Louis (2011) and the predictive capability of Blos, da Silva, and Wee (2018). In stark contrast to previous categories, all but one of the agent-based approaches use multi-agent systems as a...

    [...]

  • ...Bansal et al. (2005) propose a similar agent-based approach that monitors key performance indicators (KPIs) to identify disruptions, then evaluates corrective actions and finds the optimal one. Note that these works are the first approaches in this survey so far to offer a holistic approach. Multi-agent systems have also been used as simulation tools for inventory management to reduce costs and improve fill rate (Chan and Chan 2006), to resolve collaboration issues among supply chain entities that arise due to uncertain supply and demand (Kwon, Im, and Lee 2007) and to reduce costs and the bullwhip effect in a multi-stage supply chain (Zarandi, Pourakbar, and Turksen 2008). In all the aforementioned approaches, agent technology is preferred over mathematical optimisation due to its inherent capability to capture negotiation and coordination among different parties and the relatively less computational effort required. In contrast, Mele et al. (2007) propose a hybrid approach for retrofitting a production/distribution supply chain design to address uncertain demand, transport and processing times....

    [...]

  • ...Bansal et al. (2005) propose a similar agent-based approach that monitors key performance indicators (KPIs) to identify disruptions, then evaluates corrective actions and finds the optimal one. Note that these works are the first approaches in this survey so far to offer a holistic approach. Multi-agent systems have also been used as simulation tools for inventory management to reduce costs and improve fill rate (Chan and Chan 2006), to resolve collaboration issues among supply chain entities that arise due to uncertain supply and demand (Kwon, Im, and Lee 2007) and to reduce costs and the bullwhip effect in a multi-stage supply chain (Zarandi, Pourakbar, and Turksen 2008). In all the aforementioned approaches, agent technology is preferred over mathematical optimisation due to its inherent capability to capture negotiation and coordination among different parties and the relatively less computational effort required. In contrast, Mele et al. (2007) propose a hybrid approach for retrofitting a production/distribution supply chain design to address uncertain demand, transport and processing times. A genetic algorithm is employed to solve a multi-stage stochastic model of the problem, but the fitness of each iteration’s individuals is calculated using agent-based simulation. The work of Giannakis and Louis (2011) is notable as it goes beyond simulation to explore the learning capabilities of multi-agent systems in the context of SCRM. The proposed framework addresses all SCRM phases and starts similarly to others by monitoring KPIs to identify deviations, which are then assessed in order to pinpoint the root cause. Through case-based learning, successful past decisions are leveraged when similarities with the current situation are identified (e.g. with regard to root causes). The optimal decision is selected and its results simulated and quantified, with all the information stored in order to be considered in similar situations in the future. Recent work by Blos, da Silva, and Wee (2018) proposes a holistic disruption management framework that combines multi-agent systems with variants of Petri Nets called Production Flow Systems and Place-Transition Petri Nets. The latter are used to model complex supply chain networks in several disruption scenarios. An agent-based system is used to generate more scenarios based on the existing ones, choose the most appropriate mitigation actions and evaluate the resulting effect in performance. Then, based on a risk event database, the supply chain is monitored in order to detect disruptions and immediately propose mitigation plans. As shown in Table 3, the aforementioned agent-based approaches usually focus on very few parameters (up to 4), since it is difficult and computationally complex to simultaneously support a large number of policies that formalise agent behaviour. Similarly to previous categories, large datasets and predictive or learning capabilities are not supported, with the exceptions of the learning capability of Giannakis and Louis (2011) and the predictive capability of Blos, da Silva, and Wee (2018)....

    [...]

  • ...Bansal et al. (2005) propose a similar agent-based approach that monitors key performance indicators (KPIs) to identify disruptions, then evaluates corrective actions and finds the optimal one. Note that these works are the first approaches in this survey so far to offer a holistic approach. Multi-agent systems have also been used as simulation tools for inventory management to reduce costs and improve fill rate (Chan and Chan 2006), to resolve collaboration issues among supply chain entities that arise due to uncertain supply and demand (Kwon, Im, and Lee 2007) and to reduce costs and the bullwhip effect in a multi-stage supply chain (Zarandi, Pourakbar, and Turksen 2008). In all the aforementioned approaches, agent technology is preferred over mathematical optimisation due to its inherent capability to capture negotiation and coordination among different parties and the relatively less computational effort required. In contrast, Mele et al. (2007) propose a hybrid approach for retrofitting a production/distribution supply chain design to address uncertain demand, transport and processing times. A genetic algorithm is employed to solve a multi-stage stochastic model of the problem, but the fitness of each iteration’s individuals is calculated using agent-based simulation. The work of Giannakis and Louis (2011) is notable as it goes beyond simulation to explore the learning capabilities of multi-agent systems in the context of SCRM. The proposed framework addresses all SCRM phases and starts similarly to others by monitoring KPIs to identify deviations, which are then assessed in order to pinpoint the root cause. Through case-based learning, successful past decisions are leveraged when similarities with the current situation are identified (e.g. with regard to root causes). The optimal decision is selected and its results simulated and quantified, with all the information stored in order to be considered in similar situations in the future. Recent work by Blos, da Silva, and Wee (2018) proposes a holistic disruption management framework that combines multi-agent systems with variants of Petri Nets called Production Flow Systems and Place-Transition Petri Nets....

    [...]

Journal ArticleDOI
TL;DR: In this article, the authors developed a framework for the design of a multi-agent based decision support system for the management disruptions and mitigation of risks in manufacturing supply chains, which is based on the use of modern information technology (IT) decision support systems.

321 citations

Journal ArticleDOI
TL;DR: In this article, a heuristic rescheduling strategy is proposed that overcomes the shortcomings of existing approaches for generating (near) optimal schedules for a real-world refinery typically require significantly large amounts of time.
Abstract: Globalization and the resultant complexity of today's supply chains require that enterprises be agile and proactive. This communication looks at a refinery supply chain where disruptions such as crude arrival delay could make the current schedule infeasible and necessitate rescheduling of operations. Existing approaches for generating (near) optimal schedules for a real-world refinery typically require significantly large amounts of time. This is undesirable when rectification decisions need to be made in a short time. Further, when the problem data given to the existing scheduling approaches are changed, as is the case during a disruption, the optimizers follow different solution paths and result in substantially different schedules. A heuristic rescheduling strategy is proposed that overcomes both these shortcomings. The key insight exploited here is that any schedule can be broken into operation blocks. Rescheduling is performed by modifying these blocks in the original schedule using simple heuristics to generate a new schedule that is feasible for the new problem data. Our strategy avoids major operational changes by preserving—as far as possible—the blocks in the original schedule. The major advantages of the proposed method are its real-time computational performance and the minimal changes to the operations as compared to total rescheduling. Further, the proposed strategy can also identify many feasible schedules and allow refinery personnel to select one by considering other factors that cannot be adequately modeled in a scheduler. Our method is illustrated using five types of disruptions occurring in a refinery. The various factors that affect the robustness of a supply chain in the face of disruptions are also discussed. © 2007 American Institute of Chemical Engineers AIChE J, 2007.

58 citations

Journal ArticleDOI
TL;DR: This work provides a unified model that integrates the subsurface, wells, and surface levels of an upstream production project and is the first contribution that uses mathematical programming in a real dynamic sense by honoring the constituent partial differential equations.

56 citations

Journal ArticleDOI
TL;DR: In this article, the authors propose stochastic dynamic decision-making tools that could be employed for capturing the trade-offs among multiple and conflicting-in-nature criteria so as to provide a design of a resilient shock absorber (RSA) for disrupted supply chain network (SCN).
Abstract: This article interweaves the widely published empirical frameworks with a new paradigm proposing stochastic dynamic decision-making tools that could be employed for capturing the trade-offs among multiple and conflicting-in-nature criteria so as to provide a design of a resilient shock absorber (RSA) for disrupted supply chain network (SCN). Modern SCNs encounter ‘excursion events’ of different kinds mainly due to uncertain and turbulent markets, catastrophes, accidents, industrial disputes/strikes in organisations, terrorism and asymmetric information. An ‘excursion event’ is an unpredictable event that effectively shuts down or has a relatively large negative impact on the performance of at least one member of a system for a relatively long amount of time. In this article, design of an analytical framework has been conceptualised that allows an SCN to avoid propagating the ill effects of the ‘excursion events’ further and maintains the network at a desired equilibrium level. A broad analytical view of e...

38 citations


Cites background or methods from "An Online Decision Support Framewor..."

  • ...Bansal et al. (2005) – Suggest a framework with an agent-based online decision support....

    [...]

  • ...Bansal et al. (2005) – Suggest a framework with an agent-based online decision support. Lehmann and Travers (2007) – Empirically analyse risks during disruption events through simulation, modelling and engineering assessment. Zhao et al. (2008) – A closed-loop SCN is studied; analytically, the coordination of the network with one supplier, multiple retailers and third-party logistics under normal and disrupted conditions is analysed. Oke and Gopalakrishnan (2009) – Empirically investigate mitigation strategies to deal with supply chain risks, identify generic strategies to handle risks and specific strategies for handling particular risks. Zhao et al. (2009) – Demonstrate feasibility of the disruptioncoordination model for a closed-loop SCN formulated in Zhao et al....

    [...]

  • ...Bansal et al. (2005) – Suggest a framework with an agent-based online decision support. Lehmann and Travers (2007) – Empirically analyse risks during disruption events through simulation, modelling and engineering assessment. Zhao et al. (2008) – A closed-loop SCN is studied; analytically, the coordination of the network with one supplier, multiple retailers and third-party logistics under normal and disrupted conditions is analysed. Oke and Gopalakrishnan (2009) – Empirically investigate mitigation strategies to deal with supply chain risks, identify generic strategies to handle risks and specific strategies for handling particular risks. Zhao et al. (2009) – Demonstrate feasibility of the disruptioncoordination model for a closed-loop SCN formulated in Zhao et al. (2008) with numerical simulations using MATLAB software. Svensson (2000) – Provide a conceptual framework to study the sources of disruption events; categorise the disruption events based on the empirical evidences. Serrano et al. (2007) – Provide a supply chain loss minimisation technique; Non-Dominated Sorting Genetic Algorithm for Multi-Objective Optimisation (NSGA-II) is used to manage the impact of the disruption that recuperates the supply chain losses. Handfield et al. (2007) – Empirical framework to identify and mitigate risks during the impact of disruption events through a series of focused interviews with senior executives engaged in managing supply chain risks. Yu et al. (2009) – Evaluate the impacts of disruption risks in SCN in a two-stage supply chain with a non-stationary and price-sensitive demand....

    [...]

  • ...Bansal et al. (2005) – Suggest a framework with an agent-based online decision support. Lehmann and Travers (2007) – Empirically analyse risks during disruption events through simulation, modelling and engineering assessment. Zhao et al. (2008) – A closed-loop SCN is studied; analytically, the coordination of the network with one supplier, multiple retailers and third-party logistics under normal and disrupted conditions is analysed. Oke and Gopalakrishnan (2009) – Empirically investigate mitigation strategies to deal with supply chain risks, identify generic strategies to handle risks and specific strategies for handling particular risks. Zhao et al. (2009) – Demonstrate feasibility of the disruptioncoordination model for a closed-loop SCN formulated in Zhao et al. (2008) with numerical simulations using MATLAB software. Svensson (2000) – Provide a conceptual framework to study the sources of disruption events; categorise the disruption events based on the empirical evidences. Serrano et al. (2007) – Provide a supply chain loss minimisation technique; Non-Dominated Sorting Genetic Algorithm for Multi-Objective Optimisation (NSGA-II) is used to manage the impact of the disruption that recuperates the supply chain losses....

    [...]

  • ...Bansal et al. (2005) – Suggest a framework with an agent-based online decision support. Lehmann and Travers (2007) – Empirically analyse risks during disruption events through simulation, modelling and engineering assessment....

    [...]

References
More filters
Journal ArticleDOI
TL;DR: In this article, a unified framework for modeling, monitoring and management of supply chains is proposed, which integrates the various elements of the supply chain such as enterprises, their production processes, the associated business data and knowledge and represents them in a unified, intelligent and object-oriented fashion.

234 citations


"An Online Decision Support Framewor..." refers background in this paper

  • ...Unhindered and timely material, information and finance flow between different entities of supply chain is another important element....

    [...]

Journal ArticleDOI
TL;DR: In this article, an agent-based framework for supply chain decision support systems (DSSs) is proposed to integrate all the decision-making processes of a refinery, to interface with other systems in place, to incorporate dynamic data from various sources and to assist different departments concurrently.

132 citations

01 Jan 2003

68 citations


"An Online Decision Support Framewor..." refers background in this paper

  • ...In the face of highly competitive global markets, companies are pressurized to reduce costs and increase efficiency....

    [...]

  • ...Sheffi et al. (2003) describe mechanisms which companies follow to assess terrorism-related risks, protect the supply chain from those risks and attain resilience....

    [...]

Proceedings Article
27 Sep 2004

25 citations