An evaluation of project completion with application of fuzzy set theory
The project management contains such elements as management of time, cost, communications, procurement, quality, risk or scope of project. Each of these fields can be considered as a set of constraints, and then there is a possibility to verify their fulfillment in sense of an enterprise's constraints and its environment. These constraints determine a completion of project activities and its success or failure, finally. The paper aims to present a problem of project management in terms of fuzzy constraints satisfaction problem, and then the using of constraint programming techniques to the evaluation of project completion. A fuzzy constraints satisfaction problem enables a description of data in distinct, as well as imprecise form, in a unified framework. It seems especially important in case of unique activities of project, when their estimation is based on linguistic information from experts.
The paper presents the idea of reference model of project prototyping problem for the projects that are at risk of failure. The hierarchical structure of declarative model connects two fields: functionalities of a typical service enterprise and management system of project execution in the enterprise. The functionalities as separate Constraints Satisfaction Problems (CSP) are described. CSP contains the sets of decision variables, their domains and constraints, which link these variables. The separated problems described as CSP, then in single main CSP are integrated. On the other hand, these problems can decompose into the subproblems concerning the functionalities of different fields. The open structure of model enables to solve the decision problems with different level of specificity. The decision problem can regard a query about the results of proposed decisions as well as the decisions guaranteeing the expected results. A declarative kind of proposed reference model in a natural way allows to implement its in constraint programming languages. The possibility of this approach illustrates an example.
This paper is concerned with estimating cost of various new product development phases with the use of computational intelligence techniques such as neural networks and fuzzy neural system. Companies tend to develop many new products simultaneously and a limited project budget imposes the selection of the most promising new product development projects. The evaluation of new product projects requires cost estimation. The model of cost estimation contains product design, prototype manufacturing and testing, and it is specified in terms of a constraint satisfaction problem. The illustrative example presents comparative analysis of estimating product development cost using computational intelligence techniques and multiple regression model.
Nowadays, more and more enterprises are using Enterprise Resource Planning (EPR) systems that can also be used to plan and control the development of new products. In order to obtain a project schedule, certain parameters (e.g. duration) have to be specified in an ERP system. These parameters can be defined by the employees according to their knowledge, or can be estimated on the basis of data from previously completed projects. This paper investigates using an ERP database to identify those variables that have a significant influence on the duration of a project phase. In the paper, a model of knowledge discovery from an ERP database is proposed. The presented method contains four stages of the knowledge discovery process such as data selection, data transformation, data mining and interpretation of patterns in the context of new product development. Among data mining techniques, a fuzzy neural system is chosen to seek relationships on the basis of data from completed projects stored in an ERP system.
The paper presents identifying success factors in new product development and selecting new product portfolio. The critical success factors are identified on the basis of an enterprise system, including the fields of project management, marketing and customer’s comments concerning the previous products. The model of measuring the success of a product includes the indicators such as duration and cost of product development, and net profit from a product. The proposed methodology is based on identification of the relationships between product success and project environment parameters with the use of artificial neural networks and fuzzy neural system that is compared with the results from linear model. The presented method contains the stages of knowledge discovery process such as data selection, data preprocessing, and data mining in the context of an enterprise resource planning system database. The illustrative example enhances a performance comparison of intelligent systems in the context of data preprocessing.