Tank lorry scheduling is fundamental in the management of oil delivery facilities. A set of lorries is scheduled in a cost effective way to meet oil demands from a number of gas stations. The complexity of the scheduling arises from the diversity of constraints on oil delivery environments. The paper studies a lorry scheduling system alternative to a human scheduler taking an example of actual oil delivery facilities. Recent development in artificial intelligence leads to the use of knowledge based techniques : A method so called filtered beam search is applied to reduce combinatorially explosive search spaces. A number of combinations are omitted from the solution using the fact that most gas stations are scattered along with main roads. Another method which is used for multiattribute decision analysis is applied to evaluating candidates for a feasible solution. Each is characterized by the following four attributes : vacant lorry space, assigned delivery time, assigned lorry type, and the number of lorry hatches. A prototype system is developed for the lorry scheduling in the actual oil delivery facility. The schedules produced by the system are compared with those created by the human scheduler to verify the usefulness of the system.
This paper discusses optimal flexible cyclic scheduling for a two-machine robotic cell with finite buffer for WIP's (work-in-process) such as FMC's (flexible manufacturing cells), where jobs are processed on two machines in the same order, and sent between machines by a transportation robot. Each cycle is allowed to have different types of jobs, and the objective is to find an optimal schedule that minimizes the cycle times. In this paper we propose a heuristic algorithm based on Gilmore-Gomory and Johnson methods for this problem, and show by numerical experiments that it gives good approximate schedules of about 2% or less relative errors for any size of buffer capacity in short computational time. Also, we numerically show how the system parameters affect the system efficiency, and conclude that one of the effective ways for improving the efficiency is to reduce the diversity of job processing times rather than to have large buffer and a fast transportation robot.
This paper deals with assigning N operations on parallel H machines. The earliest start time, the latest finish time and the processing time are defined for each operation. The problem is to minimize the deviation of each processing time from the desirable time interval. We propose two autonomous decentralized scheduling algorithms to solve this problem. In these algorithms, an operation is transferred from one machine to another machine on the basis of information exchanges and cooperation among subsystems. By examining numerical results, a comparison is made between two algorithms from the viewpoints of accuracy and computation time. Consequently, abetter algorithm is found to be the one based on negotiations of one subsystem with the others.
In this paper, we take up two of the most serious technical problems aiming at the practical workflow implementation, namely, how to build up the model of workflow and how to set up and solve the optimum work-schedule problem. First, for precise dynamic modeling of workflow, we establish the timed Petri Net models which are converted from state transition models based on object-oriented modeling technique (OMT). Then, we define the optimum work scheduling problem on those timed Petri Net models and develop the solution of that problem depending on linear programming techniques. Further, we set up supplemental constraints for the linear programming to avoid resource conflicts between objects in workflow. Finally, we present the numerical calculation results of our practical studies in order to prove the feasibility and efficiency of our solution.
In this paper, we consider a bicriteria two-machine flow-shop scheduling problem which is to simultaneously minimize both Cmax and Tmax criteria, and we propose a new genetic algorithm for enumerating all nondominated solutions for this problem. Our algorithm is constructed by incorporating several existing strategies developed for multi-objective optimization problems together with a new strategy which we call seeding strategy that seeds some good individuals in the initial population. We perform computational experiments in order to compare our algorithm with other existing methods.We observe from our computational results that our algorithm produces nondominated solutions very close to exact ones, and outperforms other existing heuristics.
This paper considers the production scheduling of a flexible transfer line which produces parts for multiple agricultural machinery models. The schedule must be planned so that the parts are supplied efficiently to the assembly line in the production plant. The scheduling problem in this study is divided into lot sizing and lot sequencing problems; the lot sizing problem determines the standard lot for each part type and the lot sequencing problem determines the production sequence of standard lots. In this paper, we propose an algorithm combining heuristics and SA method to solve the lot sequencing problem and describe a system developed to support the daily planning operations.
One of the predominant characteristics of chemical batch processes is that the material leaving a batch unit is fluid. Therefore, the starting times of jobs at each unit must be determined by taking into account the availability of storage between two units. In this paper, for such chemical processes, a scheduling algorithm using simulated annealing (SA) method is proposed. In the proposed method called repetitive SA method, in order to reduce the probability of being trapped in a : bad local optimum, the scheduling using SA method is executed repeatedly, and a best schedule is selected at the final stage of repetition. The problem is how to determine the number of repetitions of scheduling and the number of schedules searched in a round of scheduling. In order to find out the best combination of those two numbers, first the probability distributions of the performances of the schedules are calculated for various cases where the number of schedules searched in a round of scheduling is different from one another. Then, the best combination of those numbers is selected using the derived probability distributions. The results of applying the repetitive SA method to scheduling problems of chemical processes suggest that the proposed method is effective typically in reducing the deviation of the performances of the schedules.