Optimising the operation of cross-dock centres from short-term and mid-term perspective

Distribution centres in the supply chain are an intermediate point for fast and cost-effective distribution. For a supply chain to be efficient and effective, organisations have to use dynamic distribution centres. Cross-docking as a JIT and lean concept has attracted the attention of many researchers and practitioners. Many companies implement cross-docking because this approach helps companies reduce their cost via reducing inventory. Operations in cross-dock centres (CDCs) are influenced by the dynamic nature of the business environment, which is full of uncertainties such as unknown arrival time of trucks, unavailability of resources, fluctuation in demand and supply, processing time, and unavailability of trucks. Accordingly, the main aim of this research is to minimise the impact of uncertainties on CDCs operations. Scheduling of trucks is one of the most important operational decisions for managing the daily activities of the cross-dock centre (CDC). One of the main factors that disrupts the scheduling of trucks is uncertainty in arrival time of trucks. Preparing action plans against early and delays in arrival time can reduce this negative impact on the daily operation and productivity of CDCs. While arrival time uncertainty continues to be one of the major concerns of industry, most scholars disregard the uncertainty of arrival caused by traffic jams and engine failures. Consequently, operation managers must reschedule the operation plan to prevent delays in the delivery of products and avoid possible bottlenecks caused by limited resources in the CDC. This research developed a new stochastic truck scheduling model that considers truck arrival time under three different scenarios, which are early, on time and late for both inbound and outbound trucks. The objective of this model was to minimise the effects of uncertain arrival time on total earliness and tardiness of trucks under different scenarios. To investigate the performance of the model, two sets of the problem (small- and large-scale instances) were designed. Two Meta-heuristic algorithms (PSO and GA) were developed to examine the performance of the model on solving large-scale instances. Arrival time uncertainty is the main external factor that disrupt the internal operation of cross-dock centres and can cause congestion in temporary storage and lack of resources. Therefore, it is important to consider the temporary storage and manual handling resources limited to be prepared for possible shortage in space and handling resources caused by uncertainties. Accordingly, a Mixed-Integer Programming (MIP) was developed to address limitation of storage space and resources in cross-dock scheduling. A sensitivity analysis was conducted to investigate the impact of temporary storage in zero, limited, and unlimited scenarios. The above two models specifically focused on daily operational planning. Cross-docking includes short-term, mid-term, and long-term decisions. Decisions made in cross-docking systems are often for a single period mainly due to the computational and operational complexity of the strategy. For this, preceding studies attempted to simplify the problems by analysing short-term problems and limiting the physical characteristics of cross-docks, e.g. a single door. This study developed a model to allow planning and allocation of inbound and outbound trucks to multiple doors across multiple periods. Accordingly, a comprehensive mixed integer programming model was presented to consider mid-term planning in the presence of a load balancing constraint to manage the internal operation. Branch-and-bound (B&B) is applied as an exact method to solve the model in small scale. Two heuristic methods PSO and GA are used in large scale instances. The results show that the genetic algorithm was highly efficient in terms of solution time and achieving the objective function. The optimality gap of the genetic algorithm was less than 2%, indicating the good performance of this algorithm.