Narrow Results

This paper concentrates on the single-machine scheduling problem with past-sequence-dependent setup times and general truncated learning effects, where the job processing times are non-increasing function of their positions in a sequence. Under common, slack and different (unrestricted) due-date assignments, our goal is to minimize the weighted sum of number of early/tardy jobs and due-date assignment cost, where the weight is not related to the job but to a position, i.e., the position-dependent weight. Under the three due-date assignments, some optimal properties and three optimal solution algorithms are proposed to solve these problems, respectively.

High volatility in customer demand orders during peak and off-peak periods is a great challenge for instant delivery. In this paper, considering the rider familiarity with different areas and the learning effect, we establish two models for different rider assignment strategies: Maximum efficiency model during the peak period and Training familiarity model during the off-peak period. Meanwhile, a hybrid algorithm parallel genetic algorithm and a large-scale neighborhood search (PGA-LNS) is designed to solve the models. The results of two comparative experiments and 50-cycle peak and off-peak alternating experiments show that adopting the Maximum efficiency model in the peak period and the Training familiarity model in the off-peak period is beneficial for instant delivery to achieve overall flexibility, stability, and delivery efficiency.