Genetic algorithm (GA) is a probabilistic algorithm for solving optimization problems, which is modeled on a genetic evolution process in biology, and is focused as an effective algorithm to find the global optimal solutions for many types of problems. Most research results on GAs are related to the design for some specific GAs and the simulation-based study for their applicability. It has not been established yet how to evaluate the performance of specific GAs in a general framework. In this paper, we give two theoretical results on the performance evaluation of GAs by using the Markovian analysis. First, we discuss convergence properties of GAs, and derive an associated performance measure, called the relaxation time for the Markov chain. Secondly, we develop an alternative Markovian analysis with rewards for real-coded GAs having real-vector individuals.
Software engineering organizations now tend to improve their processes according to such standards as CMM (Capability Maturity Model), a process evaluation model, in addition to improving their software products. This is because it is believed that mature engineering processes can bring high-quality products within the original schedule. It is reported, however, that the result of the improvement activities is strongly related with engineer satisfaction for the process, that is, effective improvement can be expected on the phases where engineers have not been satisfied with. This paper proposes a method to evaluate how much engineers feel the current engineering process needs to be improved, with Descriptive AHP (Descriptive Analytic Hierarchy Process), one of decision making tools. Our method, based on the idea that the necessity evaluation can be related with engineers satisfaction, evaluates four engineering phases under six criteria defined at CMM Level 2, determining which phases should be improved first. As a result of having system engineers use this method, we found that quality assurance activities in the manufacturing and examination phases need to be superior for improvement to other phases.
It is reported that the CNNs (Cellular Neural Networks) have a high ability of associative memory. However, the number of memory patterns and their similarity influence the recall capability of conventional CNNs. Hence, it is considered that the appropriate number of memory patterns which gives the maximum of the recall capability exists. In this paper, we investigate the influence of memory patterns on the recall capability in CNNs, and estimate the range of the appropriate number of memory patterns. Furthermore, we propose the CNN for associative memory with Multiple Memory Tables (MTT-CNN) in order to apply to practical fields. Thereupon, in order to verify the usefulness of the MMT-CNN, we performed the experiments by using 6×6, 12×12, 15×15 MMT-CNN, and examined the incomplete recall rate in each experiment. As a result, the incomplete recall rate of the MTT-CNN was far below that of the conventional CNN, and we could confirm an improvement of self-recall function. Moreover, we could avoid the incomplete recall completely by choosing more appropriate divisions.
This paper describes the recognition method of operation with hydraulic excavators. Features of operation are extracted from the lever operating data to recognize the operation. Using fuzzy reasoning depending on these features, high accuracy recognition became possible for various operators. We have also developed tuning method for membership functions in fuzzy reasoning. Even if the type of hydraulic excavators is changed, high accuracy recognition is maintained by this method. These methods were experimentally shown to be useful and have been put to practical use.
This paper deals with simultaneous determination of ordering and scheduling in a manufacturing process. The objectives of this problem are to minimize both the out-of-stock ratio and the total inventory. We propose an iterative algorithm which solves successively the respective problems for the inventory control system and the manufacturing system. Moreover, a new ordering policy is proposed in the inventory control system. In this policy, the ordering amount is determined in such a way that the minimum of margin stock ratio in all items is maximized. In the manufacturing system, the schedule is determined for that order. It is confirmed from the computational result that the proposed method outperforms the method without iteration and a periodic inventory based method.
This paper proposes model predictive control (MPC) design method for the max-plus linear systems with adjustable parameters. The several studies have been done so far under the constraints that system parameters are constant. However, it is common and inevitable that system parameters are changed in practical systems, so, it is worthwhile to design a new MPC controller with variable parameters. In this paper, we handle a max-plus linear system with system parameters dependent upon event counters and the optimization method we proposed recenlty for an optimal inverse system of max-plus linear system with linear-parameter-varying structure is applied to MPC. Since MPC uses greater prediction step than that of the inverse systems, the controller can be made more robust and the operation cost can be reduced. Finally, the proposed method is applied to a production system with three machines, and the effectiveness of this method is verified through numerical simulations.
A parallel implementation of a primal-dual interior point method is presented for multi-period and multi-facility production planning problems with nonlinear constraints. Exploiting the nested primal block-angular structure of the problem, the system of linear equations for calculating the search direction is decomposed into independent systems of linear equations and another system of linear equations which can be solved efficiently owing to the staircase structure of the coupling constraints. Computational experiments on the parallel computer VPP800 shows that the proposed method is very efficient for large-scale multi-period and multi-facility production planning problems.