A fuzzy model for generalized transportation problems in China
Mohanad R. Aljanabi,Keivan Borna,Shamsollah Ghanbari
DOI: https://doi.org/10.1002/cpe.8126
2024-04-26
Concurrency and Computation Practice and Experience
Abstract:Summary This study underscores the growing significance of multimodal transportation within the cargo sector and its consequential environmental impacts. We present a novel mathematical model for operation scheduling, incorporating variables such as resource availability, customer service benchmarks, and environmental considerations. Our objective is to mitigate transportation expenses and reduce delivery delays. The proposed approach advocates LU decomposition with a pivot strategy for rapid model resolution, adherence to convergence criteria, optimization of cost strategies, and efficient resource utilization. Leveraging the adaptive neural fuzzy inference system (ANFIS) and genetic algorithm (GA), our methodology facilitates learning from past decisions to enhance solutions, aligning supply, and demand efficiently. We evaluate financial and environmental implications across four scenarios, offering insights into the economic and environmental advantages of various transportation modes—trains, ships, airplanes—compared to truck transportation, with a specific focus on CO2 emission impacts. Implementing the ANFIS+GA model in multimodal scenarios yield impressive results: minimal MAPE transportation cost of 0.17%, R2 transportation cost of 0.996, MAPE CO2 emissions of 0.13%, and R2 CO2 emissions of 0.996. By identifying cost‐efficient routes and optimizing resource allocations, our approach enables informed decisions regarding vehicle distribution, supplier selection, and contract negotiations. Additionally, we use LU decomposition to establish the supplier risk threshold, crucial for comparing emission trade variances. Multimodal scenarios typically yield lower emissions, favoring buying emission allowances low and selling them high. Notably, the risk threshold affects low‐emission provider utilization, impacting transportation emissions. With a risk threshold of 0.12 and an emission price of 1.2, our ANFIS+GA‐based multimodal approach achieves a significant −20% deviation in CO2 emissions.
computer science, theory & methods, software engineering