International Journal of Information Technology & Decision Making

We propose an innovative AHP-based model to assess the severity of the harms a criminal commits to society in a comprehensive and coherent way. Different from the traditional approach of structuring alternatives into one level, we organize the alternatives into multiple levels of that hierarchy. This arrangement and evaluation of alternatives differs from one criterion to another, which adds to the sophistication of the task when assessing numerous heterogeneous criminal activities. Structuring multifaceted decisions with the proposed model enables us to better deal with its inherent complexity. Our approach can systematically tackle problems of widespread orders of magnitude of criteria and alternatives. When the sizes are actually very small or very large, the accuracy of rating alternatives one at a time is very low. With the proposed method, we are able to obtain priorities with greater precision.

In recent past, it has been seen in many applications that synergism of computational intelligence techniques outperforms over an individual technique. This paper proposes a new hybrid computation model which is a novel synergism of modified evolutionary fuzzy clustering with associated neural networks. It consists of two modules: fuzzy distribution and neural classifier. In first module, mean patterns are distributed into the number of clusters based on the modified evolutionary fuzzy clustering, which leads the basis for network structure selection and learning in associated neural classifier. In second module, training and subsequent generalization is performed by the associated neural networks. The number of associated networks required in the second module will be same as the number of clusters generated in the first module. Whereas, each network contains as many output neurons as the maximum number of members assigned to each cluster. The proposed hybrid model is evaluated over wide spectrum of benchmark problems and real life biometric recognition problems even in presence of real environmental constraints such as noise and occlusion. The results indicate the efficacy of proposed method over related techniques and endeavor promising outcomes for biometric applications with noise and occlusion.

This paper aims to develop a multi-criteria decision analysis (MCDA) tool for rainwater harvesting (RWH) on a university campus in Bogota, Colombia. The specific type of MCDA our research proposes is known as CRIDE (multi-CRIteria DEcision support tool), which allows us to study various methods of RWH. In order to account for uncertainties such as the variability of expert opinions and imprecision stemming from the different criteria required, we have designed this tool to incorporate ELECTRE II and the Monte Carlo method. For this case study we apply a methodology that evaluates the performance of six RWH scenarios. These six input scenarios emerged from an ideal one, wherein all rainwater throughout the campus is successfully harvested and treated until potable. From there, reality sets in: the other five scenarios are varying degrees of deviance from the ideal one, both in terms of water quantity and quality. In regard to collection and treatment we used sustainable urban drainage systems (SUDS). As for the evaluation of each of the scenarios, we outline eight criteria, which include hydraulic performance, compatibility with the University's Master Plan and financial aspects. Our research concludes that CRIDE as a means of support during the decision-making process generates a complete view of possible solutions, in addition to ranking these solutions. The CRIDE-based approach allows for the proposal of a methodology that evaluates the relevance of an established criteria set, using robustness, redundancy and sensitivity.

The processes of generating innovative solutions mostly rely on skilled experts who are usually unavailable and their outcomes have uncertainty. Computer science and information technology are changing the innovation environment and accumulating Big Data from which a lot of knowledge is to be discovered. However, it is a rather nebulous area and there still remain several challenging problems to integrate the multi-information and rough knowledge effectively to support the process of innovation. Based on the new cross discipline Extenics, the authors have presented a collaborative innovation model in the context of Big Data. The model has two mutual paths, one to transform collected data into an information tree in a uniform basic-element format and another to discover knowledge by data mining, save the rules in a knowledge base, and then explore the innovation paths and solutions by a formularized model based on Extenics. Finally, all possible solutions are scored and selected by 3D-dependent function. The model which integrates different departments to put forward the innovation solutions is proved valuable for a user of the Big Data by a practical innovation case in management.

This paper employs a real options approach to analyze optimal investment decisions. When investment projects have the characteristics of irreversibility, uncertainty and the option to wait or exit, the traditional net present value (NPV) method would underestimate the value of investment, since it neglects the values of timing and operational flexibility. The distinctive feature of this paper is that the effects of product life cycle (PLC) as well as market power are incorporated into the model. In addition, and different to the approach in Liao et al. [Optimal investment decision and product life cycle: A real options approach, Sun Yat-Sen Management Review 11(3) (2003) 1–36], we introduce the concept of technological innovation into the model. It is shown that the optimal waiting time for the investment is longer than both those in the American call options model of McDonald and Siegel [The value of waiting to invest, Quarterly Journal of Economics 101(4) (1986) 707–727], which does not incorporate dividend yield, and Liao et al. [Optimal investment decision and product life cycle: A real options approach, Sun Yat-Sen Management Review 11(3) (2003) 1–36], but is shorter than that in Dixit and Pindyck's [Investment under Uncertainty (Princeton University Press, Princeton, NJ, 1994)] model, which incorporates dividend yield. Finally, a comparative static is used to analyze the determinants of optimal investment decisions. Our results indicate that the investment-ratio threshold will be higher, and thus the optimal entry time for an investment will be delayed, when (1) the PLC is longer, (2) the uncertainty is greater, (3) the discounting rate is higher, (4) market power is larger, (5) jump size intensity is stronger and (6) the payoff out ratio (R&D/revenue) is larger.

ARACE is a new method developed within the framework of Verbal Decision Analysis (VDA). VDA methods work with verbal form of preference elicitation and evaluation of alternatives without resorting to numbers. ARACE is based on the ideas of the VDA method ZAPROS but uses a more flexible approach to inconsistency of the decision-maker's preferences by introducing a special construct of a "quasi-expert." Mismatched preferences of the decision-maker are viewed as preferences of several quasi-experts. The preference system for each quasi-expert is transitive, leading to consistent decision rules formed for each quasi-expert separately. These rules are used to compare alternatives. Differences in possible comparison of alternatives based on different quasi-experts are resolved through a Clustered Rankings Method for rankings with ties.

To enhance the prediction accuracy for crude oil price, a novel ensemble learning paradigm coupling complementary ensemble empirical mode decomposition (CEEMD) and extended extreme learning machine (EELM) is proposed. This novel method is actually an improved model under the effective "decomposition and ensemble" framework, especially for nonlinear, complex, and irregular data. In this proposed method, CEEMD, a current extension from the competitive decomposition family of empirical mode decomposition (EMD), is first applied to divide the original data (i.e., difficult task) into a number of components (i.e., relatively easy subtasks). Then, EELM, a recently developed, powerful, fast and stable intelligent learning technique, is implemented to predict all extracted components individually. Finally, these predicted results are aggregated into an ensemble result as the final prediction using simple addition ensemble method. With the crude oil spot prices of WTI and Brent as sample data, the empirical results demonstrate that the novel CEEMD-based EELM ensemble model statistically outperforms all listed benchmarks (including typical forecasting techniques and similar ensemble models with other decomposition and ensemble tools) in prediction accuracy. The results also indicate that the novel model can be used as a promising forecasting tool for complicated time series data with high volatility and irregularity.

A hybrid Particle Swarm Optimization (PSO) that features an automatic termination and better search efficiency than classical PSO is presented. The proposed method is combined with the so-called "Gene Matrix" to provide the search with a self-check in order to determine a proper termination instant. Its convergence speed and reliability are also increased by the implementation of the Principal Component Analysis (PCA) technique and the hybridization with a local search method. The proposed algorithm is denominated as "Automatically Terminated Particle Swarm Optimization with Principal Component Analysis" (AT-PSO-PCA). The computational experiments demonstrate the effectiveness of the automatic termination criteria and show that AT-PSO-PCA enhances the convergence speed, accuracy and reliability of the PSO paradigm. Furthermore, comparisons with state-of-the-art evolutionary algorithms (EA) yield competitive results even under the automatically detected termination instant.

Sales forecasting is very beneficial to most businesses. A successful business needs accurate sales forecasting to understand the market and sales trends. This paper presents a novel sales forecasting model by integrating support vector regression (SVR) and bat algorithm (BA). Since the accuracy of SVR forecasting mainly depends on SVR parameters, we use BA for tuning these parameters because Bat is a newly introduced algorithm and has many parameters. In order to find the best set of BA parameters Taguchi method was utilized. We validated our model on four known UCI datasets. Then we applied our model in printed circuit board (PCB) sales forecasting case study. We compared the accuracy of the proposed model with Genetic algorithm (GA)–SVR, particle swarm optimization (PSO)–SVR, and classic-SVR. The experimental results show that the proposed model outperforms the others. To ensure the robustness of our proposed model, sensitivity analysis was also done using our model to find out the effects of dependent variables values on sales time series.