Adaptive dynamic scheduling of single batch processing machines using a cooperative ACO–GA metaheuristic

Abstract

The goal of this study is to assess the important relationship between metaheuristic hybridization and dynamic batch scheduling serviceability performance across flow time efficiency, convergence speed, and sequence stability parameters. By analyzing the impact of a tightly coupled Hybrid ACO–GA framework on these parameters, this work illustrates how a pheromone-based “warm start” mechanism sets a better initialization benchmark for evolutionary refinement. Utilizing a hybrid optimization paradigm incorporated with rolling-horizon logic, this work examined the convergence processes and time-varying distributions of the stability metric as a function of job arrivals. Experimental validation involved a two-phase approach: a static baseline validated against MILP solvers to confirm global optimality, followed by an analysis of stochastic rescheduling agility. The empirical evidence elucidates this phenomenon and facilitates the development of more robust rescheduling strategies based on comparative magnitudes and variation patterns. In this work, a Hybrid ACO– GA model and discrete-event simulation were applied to calculate the performance distribution and its impact on the serviceability of the production environment. Hybridization has effects beyond mere efficiency. Pheromone retention is the most significant influence in the long-term stability of the model with a close Total Flow Time (TFT) gradient, which was verified in this research. This research sheds light on the crucial yet often-overlooked factor of schedule stability in dynamic batch designs. By bridging the gap between existing optimization methods and operational reality, this work offers a statistically valid (p<0.001) solution for rapid restabilization, ensuring that systems attain high throughput and predictable output over their lifetime.