Efficient Mobile Sequential Recommendation Using Backward Path Growth

The goal of mobile sequential suggestion is to give a taxi driver a route that connects a number of pick-up locations so that they are more likely to pick up clients while spending less money on gas. In essence, a significant barrier is this problem's tremendous difficulty of the computation.. In this research, we offer a unique dynamic programming-based strategy with two distinct stages-offline pre-processing as well as online search-to tackle the mobile sequential recommendation issue. From a list of pick-up spots, various candidate sequences are pre-calculated in the offline step. Based on the cost function's noted iterative feature, a backward incremental sequence creation technique is suggested. In order to efficiently minimize the search space for candidate sequences, an incremental pruning approach is simultaneously applied during the sequence generating process. The produced possible sequences are also subjected to a batch trimming technique to exclude some subpar sequences of a specified length. At the online stage, our technique can effectively select the best driving route for an empty taxi from the left candidate sequences because the pruning efficacy increases as the sequence length increases. Moreover, our approach can deal with the issue of finding the best route when there is a destination limitation or a maximum cruise distance. Experimental findings on actual and artificial data sets demonstrate that our strategy outperforms state-of-the-art approaches in regards of both pruning ability and efficiency. Hence, in practical applications, our strategies may be used to successfully handle the issue of mobile sequential recommendation with several pick-up places.