It is possible that workers' error, variation in processing time, lack of parts and machine failure will affect the delivery time of each production process. Consecutive delay in a process can lead to the postponement of production. In this paper, we propose an optimal assignment model in reset limited-cycle multiple periods to study the optimal assignment property with three work capacity levels. First, we present the reset limitedcycle model. Then, we consider an assignment with the minimized total expected cost within a certain range remove analytically determined using these three groups of workers (i.e., referred to in this paper as "local optimal assignment"). After that, we propose the theorems of the optimal assignment property. Finally, we conduct numerical experiments and discuss the optimal assignment rules withing all ranges.

When designing the variations of a product, it is necessary to make products by describing appropriate design variations and share correct information between various parties. If a common rule for descriptions is established, it contributes to the appropriate design of a product variations by understanding the number of products based on the combination of variations, or a part of the complex combination of variations. This paper proposes a description method for product variations, in order to improve design variations and compiling the manufacturing bill of materials. This method avoids too many combinations, the expansion of configuration data, and the reconstruction of that data. The proposed method is applied to some actual cases, and examined as to whether or not it is able to appropriately describe product variations. Additionally, the method is evaluated and its usefulness is shown through interviews with specialists and the performance of applied examples.

It is said that approximately 30-50% of ICT systems development have failed, therefore increasing the demand for solutions that will raise project success rates. The research aims to improve project success rates by developing a project outcome prediction system which encourages efficient participation of senior managers of projects and cross-department support members. The prediction system consists of two features — "prediction at the establishment of requirements which is an early stage of a project" and "prediction for ICT vendors which hold an important position in Japanese ICT systems development."

The prediction system was developed and evaluated in a specific ICT vendor in Japan. The data of 172 projects in the ICT vendor were assessed using a feature assessment sheet with 17 evaluation items, and a project outcome prediction model was developed by utilizing the results of the assessment and Naive Bayes classifier technique. As a result of applying 10-fold cross-validation test, the model showed 84.0% predictive accuracy. In addition, the system was evaluated utilizing ten data sets gathered from actual operations at the ICT vendor. As a consequence, 80% of the project outcomes were predicted correctly. It is thought that the prediction system will contribute to identifying projects that they should participate in preferentially.