This paper presents a new hierarchical scheduling method for large-scale production systems which include a fault tolerance mechanism based on a hierarchical Petri net model. The automobile production system equipped with two stand-by lines is focused on as one of the typical large-scale systems with fault tolerance. In the high level, the FOHPN model is used to represent continuous flow in the production of an entire system, and MLD description is used to schedule the macroscopic behavior of the production system. In the low level, TPN is used to represent the production environment of each sub-line in a decentralized manner, and the MCT algorithm is applied to find a feasible semi-optimal process sequence for each sub-line. The proposed hierarchical scheduling method is confirmed through numerical experiments.
We consider the sheet production process with one processing machine and several equipments which can be exchanged according to the variety of sheets. Equipment exchange reduces the resources, however, it causes setting up loss as well as the additional processing time. It is necessary to determine the production schedule of jobs and corresponding equipment satisfying the due date and the several requirements on sheet production. The scheduling problem should take into account some objectives such as minimizing the total processing time and loss of resources. In this paper, we propose the hybrid scheduling method by combining genetic algorithm and dynamic programming for two-objective scheduling problem, which finds Pareto optimal solutions more efficiently rather than solo genetic algorithm.
For multiple AGVs transportation systems, a collision-free routing plan of multiple AGVs is required quickly as possible. The requests for transportation are not fixed in advance in a dynamic environment. In this paper, we propose a local rescheduling procedure for a distributed and parallel route planning system in dynamic transportation. A distributed route planning system is implemented and tested on an AGV system with multiple processing systems. The effectiveness of the local rescheduling has been investigated by an experimental 5 AGVs system. The results show that the proposed rescheduling procedure can reduce 34% of total computation time compared with that of the conventional distributed route planning method while maintaining the same level of performance.
We propose a complex market-oriented programming framework based on the economics of complex systems, and develop a complex artificial market with multi-agent paradigm in this paper. Three types of heterogeneous agents are defined in the complex artificial market. It is confirmed that their interactions with micro behaviour emerge a macro order of the artificial market, and the clearing price dynamism can be analysed in economic terms. The applicability of the framework into resource allocation problem for B2B commerce is also discussed.
Recently, facility layout should be adjusted effectively to meet a dynamic change in product mix under a given system capability as product lifecycle becomes shorter. Since the total work flow depends on a selection of processing routes for all products, it is necessary in recent layout to determine not only a layout but also a set of processing routes in each period. To meet practical requirements, we should deal with different shapes and areas of facilities. Considering these three points, we propose a method in dynamic layout problem with different shapes and areas of facilities and alternative processing routes to minimize the total cost consisting of work flow costs and facility-rearrangement costs over a given planning horizon. The proposed method consists of three phases, such as a static, dynamic and modification phase. In each phase, the original layout problem is decomposed into a work flow decision problem and a layout-position decision problem. An approximate solution is obtained by solving each problem repeatedly. Some simulation experiments are demonstrated to show that the proposed method efficiently yields a near optimal layout with high accuracy.