Optimization of an Automotive Supplier's Production Line under Disruptions

Thesis Abstract

The global automotive supply chain faces recurring disruptions—from supplier bankruptcies and logistics delays to sudden demand shifts—necessitating resilient production strategies within single suppliers that operate exposed, just-in-time networks. This study addresses the problem of how an automotive component supplier can optimize its production line to maintain service levels, reduce throughput variability, and minimize total cost under disruptions such as supplier lead-time volatility, component shortages, and transportation bottlenecks. The aim is to develop an integrated decision framework that enhances line robustness while preserving efficiency in normal operation. Specific objectives include (i) diagnosing disruption sensitivity across the production line through a diagnostic model, (ii) identifying robust mix and sequencing policies that balance throughput, work-in-process, and setup costs under stochastic interruptions, (iii) designing an adaptive scheduling mechanism that responds to real-time disruption signals, (iv) evaluating the trade-offs between inventory buffers and supplier diversification as resilience levers, and (v) validating the framework on a real-world automotive supplier case with longitudinal performance data. The research adopts a mixed-methods, case-study design anchored in resilience theory and the theory of constraints, complemented by stochastic optimization and robust scheduling concepts. The population comprises a mid-size automotive component supplier operating a multi-product, multi-stage assembly line with six workstations and three critical sub-assemblies. A purposive sample of production data from 24 months, including daily demand, lot sizes, setup times, yield rates, machine downtime events, and supplier lead times, will be collected, amounting to approximately 1,000 discrete production runs. Data collection instruments include ERP extracts, line-side time studies, and semi-structured interviews with line supervisors and procurement managers to capture disruption profiles and decision heuristics. Instrument validity will be established via content validation by a panel of three manufacturing engineering experts; reliability will be assessed with Cronbach’s alpha for survey items and test-retest reliability for time-series datasets. Analytical methods combine descriptive statistics, stochastic programming, and robust optimization. A two-stage stochastic programming model with and without recourse will be developed to optimize production scheduling, lot sizing, and sequencing under scenarios reflecting lead-time variability, demand volatility, and component shortages. A robust counterpart formulation will be employed to identify schedules with minimum worst-case cost and service-level breaches. The framework will be extended with a discrete-event simulation to capture shop-floor dynamics and verify model outputs under realistic operating conditions. Regression analyses will quantify the impact of disruption indicators on key performance metrics, while ANOVA will test differences in performance across policy variants. The analysis will incorporate process capability indices and queuing theory metrics to examine throughput, WIP, and cycle time under stress. A thematic analysis of interview data will identify managerial levers and organizational factors that influence resilience, mapped to the theoretical constructs of dynamic capability and the theory of constraints. Expected findings include (i) quantifiable improvements in service level by 6–12 percentage points and reductions in average late-time by 8–15%, (ii) a robust policy set that reduces total cost variability by 10–20% under disruption scenarios, (iii) demonstrable gains in line throughput stability through adaptive sequencing tied to real-time disruption signals, and (iv) evidence that a moderate strategic buffer of critical components and supplier diversification yields higher resilience gains than excessive buffer inventories. The study contributes to knowledge by integrating stochastic and robust optimization with discrete-event simulation in a real-world automotive supply context, extending resilience literature in production systems, and providing a practical decision-support framework that links operational scheduling to organizational capabilities as described by dynamic capability theory. It will offer a tested protocol for implementing disruption-aware line optimization, including data requirements, model specifications, and steps for translating findings into actionable production plans. The main conclusion is that combining adaptive scheduling, strategic buffer positioning, and supplier risk diversification yields superior resilience without substantial cost penalties; recommendations include adopting a modular optimization toolkit integrated with ERP, establishing disruption signal protocols, and conducting regular stress-testing of schedules to sustain performance under evolving disruption patterns.