A generic representation of supply network resilience using simulation based experimentation

ABSTRACT

This paper presents a generic representation of supply network resilience implemented using Agent-Based Modelling and Simulation (ABMS) and simulation experiments. The representation captures the relationships between tactical and strategic decisions and their impact on the performance of multi-echelon networks under supply uncertainty. This representation expands practitioners’ understanding of the behaviour of their supply chains by using the Supply Chain Operations Reference model processes and performance metrics and developing a simulation model that can be adapted to a firm’s unique supply structure (sourcing strategy) and its product’s technological assembly order (Bill of Materials). To demonstrate the applicability and usability of the proposed representation, a case with single and dual sourcing network structures was analysed under disruptions of operational and strategic nature. The example demonstrates the flexibility and functionality of the proposed representation to understand the performance and resilience of multi-echelon networks under a wide range of disruptive scenarios.

KEYWORDS:

• Resilience
• supply chain disruptions
• agent-based modelling and simulation
• SCOR

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. For a detail analysis of product design theories and methodologies, the author suggests to review Tomiyama et al Tomiyama et al. (2009). Design methodologies: Industrial and educational applications. CIRP Annals-Manufacturing Technology, 58(2), 543–565.

2. This work considers only the case where customers demand is restricted to the final product, hence all the entries of the vector $\vec{x}$ are 0 except for the last one that represents the quantity demanded.

3. For a more detailed derivation of these results see Bunke et al. (2004). Bills of Material and Linear Algebra (Management Mathematics for European Schools -MaMaEuSch, Issue. U. Verlag.

4. Previous versions of SCOR follow a similar approach: it wasn’t until Revision 11.0 that the Enable process was elevated to a Level 1 process.

5. Lowest level in the BoM is assumed as 1 and, as levels go higher, the part level increases by 1. The part-level index is calculated as:

where $e_t$ is the number of elements in a component t and $Bo{M}_t$ is the BoM level of component i.

6. Base case: No disruptions

Scenario A: Operational disruption at the node level

Scenario B1: Operational disruption at the region level

Scenario B2: Strategic disruption at the region level

Scenario C: Operational disruption at both the node and region levels

Scenario D: Operational disruption at the node level and strategic disruption at the region level.

7. Scenario B2: Strategic disruption at the region level.