Agent-based Scheduling

Date: August 14
Time: 13:30--15:30
Room: 301A

(Note: Click title to show the abstract.)

Wang, Junqiang (Northwestern Polytechnical Univ.)

Abstract: The minisymposium is orgnized by the Scheduling Society of Operations Research Society of China (ORSC-SS).

Scheduling was originated 60 years ago, starting with solving n/2/F/Cmax problems by S. M. Johnson in 1954. Considering jobs or orders that might come from several customers who have different objectives or priority requests, a two-agent scheduling problem was introduced 14 years ago by A. Agnetis et al in 2000. The two-agent scheduling problem breaks the limitation that all jobs must meet same criteria, instead, in a two-agent case, jobs coming from different agents may have different criterion. Scheduling with multiple agents has already received considerable attention in the literature. Various results on complexity analysis and algorithms have been developed with different objective functions and different machine settings for two agent settings.

However, the existing researches do not give full play to agent¡¯s initiative. All proposed heuristic algorithms are exquisitely designed and fine tuned to satisfy each agent¡¯s constraints and goal. These are done by managers rather than agents. Agents are merely representing different job sets. However, in the literature of multi-agent systems (MAS), an agent is considered as an autonomous entity who can observe and act upon its environment, and direct its activity towards its goals. Agents can develop schedules using negotiation, competitive or other mechanisms rather than simple dispatching rule in heuristic and generate-and-test search strategy in meta-heuristic. Unfortunately, the two-agent scheduling research has largely ignored this autonomy property of agents focusing instead on optimizing schedules or analyzing complexity.

In the past 25 years, multi-agent system (MAS) is introduced into production scheduling to direct towards the support for scheduling agility rather than scheduling optimality and scheduling flexibility in the earlier research. However, the gap between current agent-based scheduling and initiative scheduling mainly reflects in these aspects: 1) independent agent for individual machine or job instead of a kind of machines or jobs according to the functional or physical decomposition approach; 2) open architecture to access to cloud manufacturing instead of closed architecture motivated by the manufacturing mode of real factory; 3) heterogeneous agents instead of homogeneous agents; and 4) agent-centered scheduling mode instead of manager-centered scheduling mode. In addition, there lacks richer literature discussing how to realize the autonomous ability in the context of production scheduling.

With the high technology development of internet of things, cyber-physical system, and cloud computing, a flatform is being realized where autonomous agents interact with each other to reach common objectives in the presence of real-time information, while simultaneously each agent pursues individual goals. Hence, how to autonomously perform reactive, proactive, and social actions in its execution environment and act upon each other to generate schedule is a worthy question to explore, especially in a dynamic manufacturing environment. For example, in job shops, inevitable and unpredictable disturbances necessitate the revision of established schedules during manufacturing processes.

In this symposium, we discuss some evolving scenarios, models, solution techniques and future research of agent-based scheduling.

Synchronization in Manufacturing Planning and Scheduling with Physical Internet
Huang, George Q. (The Univ. of Hong Kong)

Coordinating agents in multi-agent planning problems
Zhang, Yingqian (Erasmus Univ. Rotterdam)

Multi-agent scheduling cooperative game on a single parallel-batching machine
Zhao, Xiaoli (Inst. of Industrial Engineering & Logistics Optimization, Northeastern Univ., Shenyang)
Tang, Lixin (Inst. of Industrial Engineering & Logistics Optimization, Northeastern Univ.)

Job shop scheduling considering flexible routings based on analytical target cascading method
Guanghui, Zhou (School of Mechanical Engineering, Xian Jiaotong Univ.)
Li, Yanguang (School of Mechanical Engineering, Xi'an Jiaotong Univ.)


Code: Type-Date-Time-Room No.
Type : IL=Invited Lecture, SL=Special Lectures, MS=Minisymposia, IM=Industrial Minisymposia, CP=Contributed Papers, PP=Posters
Date: Mo=Monday, Tu=Tuesday, We=Wednesday, Th=Thursday, Fr=Friday
Time : A=8:30-9:30, B=10:00-11:00, C=11:10-12:10, BC=10:00-12:10, D=13:30-15:30, E=16:00-18:00, F=19:00-20:00, G=12:10-13:30, H=15:30-16:00
Room No.: TBA