Masaki Samejima

Dynamic optimization of virtual machine placement by resource usage prediction

In order to save energy with preventing the lack of the resources on servers in data centers, we address the dynamic optimization of virtual machine placement. A decision problem of the virtual machine placement has been formulated as Bin Packing problem. Conventional methods repeat to solve the problem at regular intervals. Live migration is used for changing the virtual machine placement, but some of live migrations are unnecessary; the virtual machine is migrated repeatedly between the physical servers. For the purpose of reducing the unnecessary live migrations, we propose a dynamic optimization of virtual machine placement by resource usage prediction. The proposed method predicts the future resource usage by Auto Regressive Model. The proposed method decides the virtual machine placement by solving the Bin Packing problem with the predicted resource usage.

SWOT Analysis Support Tool for Verification of Business Strategy

To verify business strategies, it is essential to extract and analyze business information. The amount of business information is so enormous that it is time-consuming for an analyst to extract and analyze. Thus, we propose a support tool for verifying business strategies using SWOT analysis. This tool supports the extraction of factors for analysis. However, the information includes unnecessary factors, same factors in different expressions, and meaningless factors in independent use. To solve these problems, we propose a framework for this tool using our proposed keywords extraction method.

Business scenario design support by qualitative-quantitative hybrid simulation

We propose a simulation method on qualitative and quantitative hybrid model. In order to evaluate factors of a model with qualitative causal relations, we introduce a statistical approach based on propagation and combination of effects of factors by Monte Carlo simulation. In propagating an effect, we divide a range of a quantitative factor by landmarks and decide an effect to a destination node based on the divided ranges. In combining effects, we decide an effect of each causal relation using effects ratio and sum all effects. Through applied results to practical models, it is confirmed that there are mostly same between results derived from quantitative relations and results derived from the proposed method at 0.05 alpha-level.

Risk-cost optimization for procurement planning in multi-tier supply chain by Pareto Local Search with relaxed acceptance criterion

We address a 2-objective optimization problem to minimize a retailer’s procurement cost and risk that is evaluated as recovery time of the retailer’s business after the procurement is suspended by a catastrophic event. In order to reduce the recovery time, the retailer needs to decentralize ordering to multiple suppliers and have contingency stock, which costs the retailer. In multi-tier supply chains, not only the retailer’s procurement plan but also their suppliers’ procurement plans affect the retailers’ risk and cost. Due to the huge combinations of their plans, it is difficult to find Pareto optimal solutions of the 2-objective optimization problem within a short space of time. We apply Pareto Local Search (PLS) based on heuristics to generate neighbors of a solution by changing suppliers’ plans in the closer tier to the retailer. The original PLS accepts the solutions that are nondominated neighbor solutions for the next search, but the acceptance criterion is too strict to find all Pareto optimal solutions. We relax the acceptance criterion in order to include dominated solutions whose Pareto rank is equal to or less than a threshold. The threshold is updated based on changes of Pareto rank during local searches.

Social consensus making support system by qualitative and quantitative hybrid simulation

This paper addresses support for making the social consensus on risk-reducing plans among experts and stakeholders. Parameters are given on risk-reducing plans in order to decide which risk-reducing plans are performed, but the parameters are so uncertain that experts can not set values to parameters. So, it is difficult for experts to set an agreed value to the parameters. Due to the uncertain parameters, if experts acquire the agreed combination of risk-reducing plans, the evaluation values are also uncertain. Therefore, it is difficult for stakeholders of risk to understand the evaluations. We propose the consensus making support system that enables experts to decide the combination by qualitative values and enables stakeholders to understand the evaluations by probability distributions. The proposed system decides the combination by converting qualitative values to quantitative values by random numbers and derive the probability distributions by Monte Carlo simulation. In order to realize these requirements, we apply the qualitative and quantitative hybrid simulation to the proposed system. As a result of the application to a consensus making problem, it is confirmed that the proposed system is effective for consensus making. And, in order to improve the system, it is necessary to support adjusting parameters to acquire the agreed combination.

IT risk management framework for business continuity by change analysis of information system

The large-scale and complicated information systems make it difficult to prevent or recover the failures by IT(Information Technology) risks. This paper reports an application of IT risk management to the business continuity. According to our survey, changes of information systems cause new IT risks. So, we propose an IT risk management framework by change analysis of information systems. By applying the proposed framework to the cases of IT risks, it has been confirmed that the proposed framework is useful to prevent IT risks for business continuity.

An Intelligent Tutoring System for Case-Based E-Learning on Project Management

A survey says that the success rate of information system development project is about 30%. In order to lead a project to a successful one, many companies performed several kinds of education methods to educate more excellent project managers. One of the education methods is the case-based learning that a learner thinks problems and solutions on the practical case. The tutor that provides advice on how to think about the case plays an important role in the case-based learning. However, tutors are fewer than learners. In order to make the case-based learning available online without the tutor, we propose an intelligent tutoring system for case-based e-Learning on project management. A tutor agent instead of a human tutor provides appropriate advice according to a learners input to the system. The proposed method compares the learners input to the answers that the human tutor preliminary decides in order to verify the learners input. A tutor agent chooses the advice corresponding to the content of the input from an advice list. The system repeats to provide advice to the learner until the learner finishes thinking about sufficient solutions.

An evaluation method of reduced procurement risks by decentralized ordering in supply chain

In supply chain, catastrophic disasters such as earthquakes may prevent procurements from suppliers, which causes losses in retailers. Decentralizing orders and urgent stocks can reduce the losses, but cost the retailer. And, the losses by the catastrophic disasters are uncertain. In order to support planning of decentralizing orders and urgent stocks, we propose the evaluation method of procurement risks that are Conditional Value at Risk (CVaR) of the losses, which are often applied for a rare event causing a heavy loss. Because the uncertain procurement risk is reduced in return for the cost, the proposed method evaluates the decentralizing orders with the urgent stock based on real option approach.

A Help Desk Support System Based on Relationship between Inquiries and Responses

We propose a help desk support system to extract FAQ (Frequently Asked Questions) automatically, to retrieve FAQ for inquiry e-mails, and to show FAQ in users’ inputting their inquiry e-mail. In the help desk, operators record inquiry e-mails and response e-mails. Between inquiries and response, there are relationship that similar words are often used. First, we propose a classification method of inquiry e-mails for describing FAQ with pairs of inquiries and responses. Second, we propose a detection method of the FAQ matching inquiry e-mail based on Jaccard coefficient between inquiries and responses. Finally, we propose a predictive search method of FAQ by matching an incomplete inquiry to FAQ.

Approximation method for chance-constrained programming of social consensus formation concerning IT risk countermeasure

The target of this paper is an integer programming with parameters of IT (Information Technology) risk countermeasures in order to decide the combination of IT risk countermeasures. Stakeholders in the social consensus formation do not always a unique value to each parameter of IT risk countermeasures. Regarding the parameters as random variables, we address chance constrained programming to minimize expectation of the objective function and to keep constraints with a certain probability. In order to solve the chance constrained programming, we apply the approximation method for converting chance constraints to deterministic constraints.