Multi-version Strategy Research Articles

Data pricing strategy based on data quality

This paper presents a bi-level mathematical programming model for the data-pricing problem that considers both data quality and data versioning strategies. Data products and data-related services differ from information products or services in terms of quality assessment methods. For this problem, we consider two aspects of data quality: (1) its multidimensionality and (2) the interaction between the dimensions. We designed a multi-version data strategy and propose a data-pricing bi-level programming model based on the data quality to maximize the profit by the owner of the data platform and the utility to consumers. A genetic algorithm was used to solve the model. The numerical solutions for the data-pricing model indicate that the multi-version strategy achieves a better market segmentation and is more profitable and feasible when the multiple dimensions of data quality are considered. These results also provide managerial guidance on data provision and data pricing for platform owners.

Consistency maintenance of Do and Undo/Redo operations in real-time collaborative bitmap editing systems

In real-time collaborative graphical editing systems, bitmap-based graphical editing systems are particularly special and practically useful ones, and Do and Undo/Redo operations are intricate problems in this field. However, existing researches on graphical editing systems are quite scanty. In this paper, based on Multi-version strategy, we propose a new approach to solve the Do and Undo/Redo consistency maintenance problems with due consideration of three possible cases: all-causal, all-independent and causal-independent-mixed operations. Compared with previous collaborative algorithms, the algorithms proposed in this paper support Do and Undo/Redo operations without requiring additional space. In addition, two example analyses are also given to prove the algorithms' effectiveness separately. Furthermore, the time complexity of the two algorithms is both O(n). Finally, a system prototype called bitmap-based Co-Graphical Editor is implemented to verify them realistically.

Consistency Maintenance of Group/Ungroup Operations in Object-based Collaborative Graphical Editing Systems

In real-time collaborative graphical editing systems, Object-based Group/Ungroup operations are frequently accessible and practically useful. However, the existing research on these operations of the graphical editing is rare and defective. In this paper, based on Multi-Version strategy and Address Space Transformation method, a new MVSDR algorithm, which is not only applied to simple operations (such as Create, Delete, ChangeATT, etc.), but also suitable for Group/Ungroup ones, is proposed to solve the consistency maintenance problem. The proposed algorithm abandons previous attempts to divide conflict operations into Real-Conflict operations and Resolvable-Conflict ones and to deal with them separately, thus making the algorithm more simple and effective. In addition, an example analysis is also given in this paper to prove the algorithm’s correctness and effectiveness.

Optimal versioning strategy for information products with behavior-based utility function of heterogeneous customers

This paper reconsiders two fundamental assumptions (i.e., on the quality of information product and the self-selection behavior of customers) that decide the optimality of versioning strategy (or vertically differentiated product line) for information products. The quality of an information product is clarified in terms of functional and nonfunctional features. The customers׳ behavior of self-selection among multiple versions of an information product is examined, and the disability of linear valuation function for exactly capturing customers׳ valuation on information products is clarified. The required quality is introduced to depict a customer׳s requirement on the quality of an information product, and a behavior-based utility function is thus defined, where a customer has a marginal decrease (different from the constant marginal in linear valuation function) on the valuation of versions with quality levels higher than the required, but the valuation of a customer diminishes quickly on versions with quality levels lower than the required. Then, a bilevel programming model is built to represent the task for optimizing the strategy of versioning an information product, with the monopolist as the leader and all customers as followers. Optimal quality levels and prices for multiple versions are obtained for a given number of versions. To deal with the nonlinearity and multimodality of this model due to correlated decision variables, a steady-state evolutionary algorithm, hybridized with the local search method (called the hybrid steady-state evolutionary algorithms), is developed to improve the global optimality of versioning schemes. Numerical experiments are conducted on the bilevel programming model with specific parameterization by using the hybrid algorithm. Experimental results verify the optimality of multi-version strategy and reveal various facets of its property. With the behavior-based valuation function of heterogeneous customers, the multi-version strategy is superior to the one-version scheme. The total profit increases logarithmically with the maximal number of versions, which means that the introduction of a new lower-quality version contributes less to the total profit of a monopolist. Furthermore, when a lower quality level version is offered to the market, higher quality versions are priced higher than they were in previous versioning schemes, indicating that a multi-version scheme can make more detailed segmentation of the market. Therefore, a monopolist is able to gain greater revenue via price discrimination on heterogeneous customers through vertically differentiating information products.

A Theoretical Basis for the Analysis of Multiversion Software Subject to Coincident Errors

Fundamental to the development of redundant software techniques (known as fault-tolerant software) is an understanding of the impact of multiple joint occurrences of errors, referred to here as coincident errors. A theoretical basis for the study of redundant software is developed which 1) provides a probabilistic framework for empirically evaluating the effectiveness of a general multiversion strategy when component versions are subject to coincident errors, and 2) permits an analytical study of the effects of these errors. An intensity function, called the intensity of coincident errors, has a central role in this analysis. This function describes the propensity of programmers to introduce design faults in such a way that software components fail together when executing in the application environment. We give a condition under which a multiversion system is a better strategy than relying on a single version and we study some differences between the coincident errors model developed here and the model that assumes independent failures of component verions.