Dynamic multi-objective balancing for online food delivery via fuzzy logic system-based supply-demand relationship identification

With the boom of Online Food Delivery (OFD) service, a large number of users and riders have joined the service system. OFD platforms need to balance the interests of multiple parties to improve service quality and retain active members. It is a challenging task with multiple optimization objectives and strong dynamism. To solve the problem, we design an identify-balance framework, which first identifies the real-time Supply-Demand Relationship (SDR) and then balances customer satisfaction and delivery efficiency based on SDR. For SDR identification, a hierarchical fuzzy logic system is designed to deal with uncertain data and generate rough labels according to domain knowledge. Besides, to compensate for the deficiency of expert experience and obtain more accurate SDR, a data-driven method based on weakly supervised learning technology is employed. For multiobjective balancing, several SDR-based strategies are proposed by dynamically adjusting the objective weights and rider distribution. The experiment results on a real-world online food delivery platform demonstrate the effectiveness and superiority of our framework.