The lexicographic bi-criteria combinatorial optimization problem to be discussed in this paper is a mathematical model of the food mixture packing performed by so-called automatic combination weighers, and it is described as follows. We are given a union I = I₁ ∪ I₂ ∪ • • • ∪ I_m of m sets of items, where for each i = 1,2,...,m, I_i = {I_ik | k = 1,2,...,n} denotes a set of n items of the i-th type and I_ik denotes the k-th item of the i-th type. Each item I_ik has an integral weight w_ik and an integral priority γ_ik. The problem asks to find a union I′ = I′₁ ∪ I′₂ ∪ • • • ∪ I′_m of m subsets of items where I′_i ⊆ I_i so that the total weight of chosen items for I′ is no less than an integral target weight T, and the sum weight of chosen items of the i-th type for I′_i is no less than an integral indispensable weight b_i. The total weight of chosen items for I′ is minimized as the primary objective, and further the total priority of chosen items for I′ is maximized as the second objective. For the case in which there are two types of items (i.e., m = 2), we propose an O(nT) time dynamic programming algorithm, applying a linear search technique. We also conduct numerical experiments to demonstrate the empirical performance.

View full abstract

Building production can be regarded as a one-piece production because not only specifications of individual buildings but also working environment for them are different. Additionally, different types of companies/enterprises participate in production. So, there are many uncertainties before production starts and disturbances occur frequently after production starts compared to production of general industrial products in a factory. Since it takes too much time to resolve problems with respect to such uncertainties/disturbances, a conventional production schedule has to include excessive slack/float. To resolve this issue, we propose a management system that consists of multiple agents and a bulletin board module to connect these agents. In the proposed system, both a job and a worker are regarded as agents. An autonomous job is referred to as a demand agent. A worker agent decomposes a job into some sub-jobs, that is, the worker agent generates new demand agents. A demand agent organizes an auction to find a worker agent which can perform a job corresponding to the demand agent itself and a worker agent bids for the demand agent to get the job. The bulletin board module, which controls the conveyance of information, realizes necessary and sufficient communications between agents. Furthermore, by introducing multi-layer simultaneous auctions, each worker agent can shorten its proposed timeline for a job, improving the accuracy of the estimation. As the result, a master schedule without excessive slack/float can be generated.

View full abstract

We consider an optimization of conflict-free routing problems for automated guided vehicles (AGV) with acceleration and deceleration. A continuous time model is developed to represent the dynamics of vehicles. In the proposed model, the transportation model is discretized into several regions. A network model is created by taking into account the acceleration and deceleration motions. The acceleration and deceleration are represented at curve locations. Column generation heuristic is used to find a near-optimal solution. In this algorithm, we construct a heuristic rule to generate a feasible solution with acceleration and deceleration of vehicles after the column generation. The pricing problem is represented by a resource constrained shortest path problem, which is effectively solved by a labeling algorithm. The dominance relation for acceleration and deceleration is addressed. In the proposed model, the dynamics of real speed AGV model are reflected into the routing problems. By comparing the performance of the conventional method, the effectiveness of the proposed method is demonstrated.

View full abstract

This study proposes a basic model for closed-loop supply chains which includes not only traditional forward supply chains for the generation of products but also reverse supply chains for the reuse and recycling of products in consideration of economic efficiency for make-to-order and remanufacturing-to-order companies. The basic model consists of four model components, i.e., clients, manufacturers, suppliers, and remanufacturers. A remanufacturer is added to the previous model of forward supply chains in this study as a new model component which collects used products from clients and provides reusable parts to manufacturers in consideration of the demand of products. Remanufacturers as well as manufacturers and suppliers modify their schedules and negotiate with each other in order to determine suitable prices and delivery times of products. Remanufacturers stimulate clients to discard used products to meet the demand of reusable parts. They can increase the amount of reused products and reduce wastes by creating a balance between supply and demand of reusable parts. A prototype of a simulation system for closed-loop supply chains is developed in order to evaluate the effectiveness of the proposed model and negotiation protocol. Experimental results of the proposed model are compared with the ones of a conventional model which discards the used products without negotiation processes between remanufacturers and clients. Experimental results show that the proposed model can reuse more products than the conventional model.

View full abstract

This paper deals with a scheduling method of manual work in consideration of On-the-Job Training (OJT). As of skillful works such as maintenance works of airplanes, workers have to learn many kinds of works with OJT. However, it is difficult to balance between working efficiency and keeping the number of skilled workers by performing OJT. The purpose of this study is to develop a method for adjusting the value of EC automatically by considering the balance between working efficiency and skill education and to find effective rules for adjusting the value of EC. To overcome this problem, the authors previously proposed an education coefficient (EC) as a parameter that adjusts the frequency of OJT and a scheduling support system for the long-term scheduling of OJT. However, it is difficult to find effective values of EC that give a good balance between working efficiency and skill education to maintain a certain number of skilled workers. In this paper, the authors propose a simulation-based scheduling support system using reinforcement learning. The objective of this system is to adjust the value of EC automatically based on the specific situations (deadline and current number of skilled workers) by considering the balance between working efficiency and skill education. This paper also shows that the learning system generates suitable schedules when the working conditions (the number of workers) change during the progress of working by using relearning methods. The experimental results show that the proposed learning system generates adequate schedules to obtain as many skilled workers as possible within the fixed time limit.

View full abstract

In this paper, we consider a combinatorial optimization problem in a cyclic production system treating identical jobs. The cyclic production system possesses m machines, and every job consists of n tasks. The n tasks must be processed in a predetermined sequence, and each task must be processed on a predetermined machine. Each machine can process at most one task at a time, and no preemption for task processing is allowed. The job may visit a machine more than once to be completed, i.e., it is a re-entrant one. We focus on a restricted cyclic schedule, saying a periodically segmental schedule, where no task lies over consecutive segments. For a periodically segmental schedule, let s denote the number of jobs being in process during a segment, we also see that the sequence of n tasks of every job is split into s disjoint subsequences of tasks. We define a widget by such a subsequence of tasks. Given a job splitting with the splitting number s, we only have to assign the s widgets on machines in a segment to obtain a periodically segmental schedule. The problem to be discussed here asks to find a job splitting with the minimum splitting number s = s^∗ so that for a given τ > 0, the generated s widgets can be assigned on machines in a segment with length τ. We propose a heuristic algorithm utilizing a dynamic programming based linear partition for finding a job splitting. Numerical examples demonstrate the behavior of the proposed heuristic algorithm.

View full abstract

This paper presents an efficient scheduling method for a job shop environment in which working shifts repeat regularly and production capacity is adjustable by adding overtime for each machine in each working shift. The primal objective of scheduling is to meet job due dates and the secondary one is to minimize total overtime. This problem is a highly complex one in which simultaneous decision making for operation sequencing and overtime usage is required to efficiently achieve the two objectives having a trade-off relationship. We have already proposed the method using genetic algorithm incorporating priority rule for this scheduling problem. In the method, priority rules for operation sequencing were the key for high performance. This paper improves the method by applying priority rules for both operation sequencing and overtime usage. For operation sequencing, a new priority rule called (SL/RPN) ^β+SPT is proposed. For the determination of overtime usage, a method using a critical-ratio-based rule is newly introduced. Numerical experiments show the effectiveness of the proposed method.

View full abstract

Recent development of computer network technologies is realizing highly-distributed manufacturing systems from the hardware point of view, where each facility is computerized and manages itself autonomously by communicating with other facilities. For this new type of manufacturing systems, a new discrete event simulation paradigm has been discussed, in which sorting of events is performed by exchanging messages about their occurrence times among the facilities. To make this paradigm beneficial enough, it is desirable that re-scheduling process performed after the simulation is carried out in a similar paradigm. For this reason, a highly-distributed scheduling method was proposed, where priority of each facility for job assignment is dynamically updated with indirect decision making and information control performed by a communication protocol. However, this method was developed only for a very simple scenario, and should be enhanced so that it can be utilized in real manufacturing. This paper presents an improved method which can handle the more complicated scenario where materials are transported to a production area one-by-one at non-constant time periods and some of them cannot be assigned to any machine at its arrival time.

View full abstract

We propose an efficient heuristic method for job-shop scheduling problems (JSP) with the objective of total weighted tardiness minimization. The proposed method uses schedule reconstructions by priority rules to guide a local search towards promising solutions. Typically priority rules determine the whole schedule and thus in corporating it with local search procedure is difficult. In our proposed method, a priority rule decides a schedule within an arbitrary selected time window and the rest of the schedule is determined by the schedule obtained by a conventional local search method. The priority rule is given by a linear combination of simple priority rules. To improve a schedule efficiently, an appropriate set of time window and gains of the linear combined priority rule is required. Therefore, we optimize the set of time window and rule gains with genetic algorithm. This rule-based reconstruction procedure and the conventional local search procedure are alternately applied to a current solution. In the experiments, the efficiency of rule-based reconstruction procedure is verified and the proposed method is compared with one of the most effective existing methods. The results show that the proposed method outperforms the existing method on large problems with sufficient computational time. The average performance is particularly improved due to the ability of rule-based reconstruction procedure to escape from local optima of the conventional local search procedure efficiently.

View full abstract