Kazuyuki Mori

Application of an immune algorithm to multi‐optimization problems

An immune system has powerful abilities such as memory, recognition, and learning to respond to invading antigens. The concept is expected to be applicable to many engineering applications. This paper proposes an optimization algorithm that imitates the immune system in order to solve multi-optimization problems partly by using a genetic algorithm. Through illustrative examples of multimodal functions such as the Shubert function, the proposed algorithm is shown to be effective in searching for a set of global solutions.

Immune Networks Using Genetic Algorithm for Adaptive Production Scheduling

Abstract Immune system has powerful abilities such as memory, recognition and learning to respond to invading antigens, and is expected to be applicable to many industrial applications as well as neural networks. This paper proposes an algorithm imitating the immune system to solve the optimization problem partly with genetic algorithm. The proposed algorithm is shown to be capable of searching for a global solution not but local solutions through an illustrative example of a resource allocation problem.

An Energy Distribution Decision Method in Distributed Energy Management Systems by the Market-Oriented Programming

Control of CO2 emissions which is the main factor of global warming is one of the most important problems in the 21st century about preservation of earth environment. Therefore, efficient supply and use of energy are indispensable. We have proposed distributed energy management systems (DEMSs), where we are to obtain optimal plans that minimize both of costs and of CO2 emissions through electrical and thermal energy trading. A DEMS consists of the plural entities that seek their own economic profits. In this paper, we propose a trading method that gives competitive equilibrium resource distribution by applying the market-oriented programming (MOP) to DEMSs

Scheduling Editor for Production Management with Human-Computer Cooperative Systems

Abstract The computer-aided scheduling environment of manufacturing system is described. The cooperation of the human and the computer is proposed, in which the complex simulation and data handling are executed by computer and the evaluation of schedule and the editing of schedule is indicated by the human planner. The cooperation of human and computer is essential in the system to edit the schedule as a planner desires. The simulation and the modelling of manufacturing systems as discrete event systems is based on the timed Petri net.

Immunity-Based Management System for a Semiconductor Production Line*

A production line of semiconductor is large scale and complex system. A control system of the line is consiEnred to be difficult to control because there exist lots of malfunctions such as maintenance of equipment, equipment break down and unbalance of WIP to disturb production of wafers in the semiconductor production system. We have been exploited some methods and systems using simulations or expert systems approach to solve these disturbances [2],[3],[4]. However, the semiconductor production systems had been large and complex and the environments of the systems have been changing dynamically, so that we could not exploit a perfect control system of semiconductor production by using only these methods. This paper presents a method applying an artificial immunity based system Enscribed by multi agent nets to adapt itself to dynamical environment

Project scheduling under partially renewable resources and resource consumption during setup operations

A RCPSP formulation that can deal with realistic energy constraints is proposed.An integer program and a constraint program of the RCPSP are presented.A method to restrict selectable modes is proposed.Results of computation evaluation reveal show effectiveness of the proposed method. There has been an increasing pressure on manufacturing industries to reduce energy consumption. In this study, we propose a new variant of RCPSP called RCPSP/πRC, which can deal with realistic energy constraints such as power restriction during peak hours, contract demand, and energy consumption during setup operations. First, we present an integer programming (IP) model and a constraint programming (CP) model of the RCPSP/πRC. Next, we present a heuristic mode restriction method called a mask calculation algorithm to achieve efficient searching by restricting selectable modes. Finally, through computational experiments, we evaluate the proposed methods and show their effectiveness.

Solving Distributed Unit Commitment Problem With Walrasian Auction

In the process of scheduling multiple generating units, an up/down decision for every generating unit has to be made for every hour on the planning horizon. Once the unit commitment is decided for every generating unit, a generation level (economic dispatch) from the committed unit is calculated to minimize the total operation cost. Such a problem is known as a unit commitment problem (UCP). Because an up/down pattern is expressed with a vector of binary variables, one has to solve a combinatorial optimization problem. Due to electricity liberalization or the introduction of distributed power sources in recent years, the need to solve the UCP by autonomous distributed agents, that is, solving distributed UCP (DUCP), comes about. This paper proposes a method of solving the DUCP with a Walrasian auction. The main feature of the proposed method is that not only a generation-level pattern but also an up/down pattern can be determined. The results of computational experiments show that the DUCP is solved through the cooperation of autonomous distributed agents by using the proposed method.

Multi-objective embarrassingly parallel search for constraint programming

Optimization plays an important role in various disciplines of engineering. Multi-objective optimization is usually characterized by a Pareto front. In large scale multi-objective optimization problems, determining an optimal Pareto front consumes large time. Thus, parallel computing is used to speed up the search. Constraint programming is one of the logic-based optimization techniques for solving combinatorial optimization problems. Kotecha et al. proposed a constraint programming-based strategy to determine an optimal Pareto front. Regin et al. proposed a parallel search for constraint programming, called the embarrassingly parallel search. In this paper, we propose the multi-objective embarrassingly parallel search for multi-objective constraint optimization, which combines the two strategies.

Risk Based Capacity Planning Method for Semiconductor Fab with Queue Time Constraints

It is difficult to plan capacities of a semiconductor wafer fabrication with constraints of queue time which means a waiting time from the end of a step to the start of the next step for warranting production quality. If the queue time constraints could not be kept, the product is scrapped or reprocessed for qualitative recovery, and production efficiency remarkably decreases. This paper proposes an evaluation method which enables capacity planning of a fab with the queue time constraints, and shows its validity by simulation experiments. Furthermore we present an effectiveness of the method by applying to actual production lines.

A Real-Time Scheduling Method Using Hierarchical Decentralized Decision Makers for Flexible Manufacturing Systems

This paper presents a hierarchical decentralized scheduling method for a real-time production line control, which consists of a strategic level scheduler(SLS) and operational level schedulers(OLS). The SLS aims at making a schedule referenced by the OLS. The schedule solved by the SLS satisfies some constraints, and is generated by a schedule editor on the Flexible Planning System(FPS). The OLS aims at the real-time control of the production line and is based on the know-how of the supervisor.