Yiwei Jiang

This article considers an identical parallel‐machine task scheduling problem motivated by operations management of online services. A task with an integer processing time can be split into sub‐tasks with integer processing times. Each task has multiple integer milestones and at each milestone a nonnegative penalty will occur. The penalty value of a task at a milestone is a convex nonincreasing function of the completed amount by this milestone. Our objective is to determine a feasible schedule for all the tasks on given identical parallel machines, such that the sum of all tasks' total penalty at all milestones is minimized. We prove the NP‐hardness of this problem in the ordinary sense and develop a branch‐and‐price algorithm. Computational experiments utilizing data from an online service operations survey show that this algorithm is singularly efficient and promising.

We consider parallel-machine scheduling in the context of shared manufacturing where each job has a machine set to which it can be assigned for processing. Such a set is called the processing set. In the shared manufacturing setting, a job can be assigned not only to certain machines for processing, but can also be processed on the remaining machines at a certain cost. Compared with traditional scheduling with job rejection, the scheduling model under study embraces the notion of sustainable manufacturing. Showing that the problem is NP-hard, we develop a fully polynomial-time approximation scheme to solve the problem when the number of machines is fixed.

We consider parallel-machine scheduling with identical machine resource capacity limits and DeJong’s learning effect. Each job has a resource consumption requirement and a normal processing time. The actual processing time of a job is a function of its normal processing time, subject to DeJong’s learning effect, while the resource consumption of a job is a function of its actual processing time. Each machine has the same resource capacity limit. The objective is to maximise the minimum machine load. Considering three resource consumption functions, namely, linear, concave, and convex, we show that all three scheduling models are NP-hard and propose two approximation algorithms for the models and analyse their worst-case ratios.