Switching resource allocation optimization method and device under multi-mode network, and medium

The invention discloses an exchange resource allocation optimization method and device under a multi-mode network and a medium, the method selects an ASIC exchange chip, an FPGA and PPK software exchange on a multi-mode network element based on machine learning, and specifically comprises the following steps: manual pre-configuration is carried out, and basic rules of multi-mode software and hardware cooperative processing are formulated; off-line learning: designing training configuration in an off-line learning stage to capture different exchange resource use variables, running an experiment to generate original data of a training classifier, and training a model off-line by using the generated performance index; and performing online reasoning to obtain exchange resource allocation suggestions, and updating the modal codes according to the exchange resource allocation suggestions. According to the method, the I-type equipment of the multi-mode network element is used, flexible and efficient allocation of t