Decision Analytica, A fuzzy MADM decision based support system

Project Title

Decision Analytica, A fuzzy MADM decision based support system

Project Area of Specialization Artificial IntelligenceProject Summary

Decision Analytica is a web application for Multi-Attributes and Multi-Criteria problems. It is a platform for Decision-Makers to provide their problem description along with the set of choices and criteria they have and get a best alternative for their problem/project. The decision-making or reasoning process for humans is complex and is affected by numerous factors. In our daily life, we mostly encounter the problems whose decisions are dependent on multiple attributes and there might be large number of choices or alternatives for that. Also, human beings are more comfortable in rating their preferences and opinions in human friendly language (i.e Very Important, Strongly Important, Less Important etc.) which creates an uncertainty in evaluating the problem. Multi-criteria decision analysis methods and algorithms are available separately for different problems and some implementations with a limited capability can be found in Excel or MATLAB codes. Moreover, the lack of any commercial tool related to this field and vast research going through many years are the prime motivating factors behind this project. Some state-of-the-art algorithms despite their advantages are not commercially available.

Decision Analytica- A Fuzzy MADM Decision-based support system finds a large scope to handle all these peculiarities and uncertainty where some useful algorithms would be set together to provide an optimal solution to decision-makers.

Project Objectives

1) Commercially launch an end-to-end platform for solving uncertain multi-criteria problems.

Separate modules in the form of some algorithmic calculations are available which needs a thorough study of Multi-Criteria Decision-Making techniques before they can actually be used. Decision Analytica will be a complete web application constituting five modules Screening, Weight Assessment, Ranking, Rank Aggregation and Sensitivity Analysis.

2) To provide a forum where an ideal solution can be provided to decision-makers without compromising an accuracy.

Another concern is the accuracy of this computational intelligence system that would provide a solution. To ensure the accuracy of every module different validation techniques will be applied. In Screening phase, Brella’s Census Evaluation or T-Test will be used. Consistency Ratio will used to check validation at Weight Assessment. Sensitivity Analysis will be applied for checking overall authentication of the system.

3) Integrate state-to-art Algorithms

Some state-of-the-art algorithms despite their advantages are not commercially available. These algorithms include Fuzzy Best Worst Method (FBWM), Ishikawa’s Fuzzy Delphi Method, Buckley’s fuzzy AHP, Fuzzy WASPAS, MULTIMOORA, Fuzzy TOPSIS, etc. Decision Analytica will be implementing some important ones from among them.

4) Providing easy-to-use website

An application with user-friendly and understandable environment which will delibrately interact with user at every stage.

5) Solving ambiguous and uncertain problems

The users feel more comfortable in expressing their opinions in linguistic terms. To address this challenge, numerous decision-making methods have been devised to assist users and decision makers. However, decision-making process is not a straightforward task, as it may involve several uncertainties, hindering the sound decision-making. Decision Analytica will handle these ambiguous problems.

Project Implementation Method

We will implement this project by following the standard development methodology of Rational Unified Process (RUP) defined by IBM, which has four constituent phases Inception, Elaboration,  Construction, and Transition. Similarly, This development methodology has a technical workflow that includes Requirements, Analysis, Design, Implementation and Test workflows. These workflows are judged on the basis of Inception Phase, Elaboration Phase, Construction Phase and Transition Phase.

This development methodology explicitly enforces compilation of UML artifacts for design & analysis including Domain model, Activity Diagram, Use case Diagram, Class Diagram, Sequence Diagram, Operations contracts.

The attached Use Case Model shows the overall interaction between the user and system. An Activity Diagram shows how the modules of system are working technically.

The project is being developed progressively keeping in view that implementation will be integrated by the end of August 2021. The following processes are completed until now:

(a) The authentication module which takes the data from the user (decision-maker) in sign up form and saves it for later login.

(b) The Information-taking process in which Alternatives (choices for the problem), Criteria, and their description is being taken from the user.

(c) The Questionnaire for Screening- The user is shown with Criteria to rate them from Extremely to Moderately Important in comparison with other Criteria.

(d) A Fuzzy Delphi Method-  The algorithm is written in Python to screened out unimportant criteria from the list and show only those criteria that are actually important to be considered for the problem. A linguistic value is obtained against each opinion of user, the value is then converted into triangular fuzzy number using concepts of minimum, maximum and geometric mean. This triangular fuzzy number is then defuzzified to obtain single value. Now this value will be compared with threshold value to screen out unimportant criteria.

The output of one module will be an input for the next module. The output achieved from screening module then will be used for weight assessment. We will use Extent Analysis Method with AHP for this. Then after weight assessment Ranking will be performed using Fuzzy TOPSIS or MULTIMOORA and Borda count will be applied for Rank Aggregation.

Finally Sensitivity Analysis will be applied to test overall responsiveness of System.

The Domain Model shows the complete business modeling of our Project.

Benefits of the Project

Decision-makers need to learn the algorithms and techniques if they want to use MADM for their Multi-Attribute problems. Utilizing the techniques separately will be costly. Moreover, the assurity over accuracy cannot be guaranteed.

1) Decision Analytica will provide an end-to-end framework for providing an optimum solution- all at one place

2) Designing first ever commertial tool to target large client base of Organizational Decsion-makers

A large Client base for Organizational Decision-Makers and Academic Researchers exist, but no generic tool is available in the market. So, we will be targeting these communities. These Organizational Decision-Makers can use Decision-Analytica to select contractors, suppliers, IT officers or products.

3) Deals uncertainty at every stage

4) Ensures accuracy at each level

5) Scalable for small to large problems

6) Targets Naive to Expert Users

Decision Analytica can be used by Academic Researchers who are confused to work in certain fields. Ranking universities problem, strategic vendor selection, ELT software selection and investigation of feasibility study for solar farms are some examples of real-life problems where MADM techniques were used.

Technical Details of Final Deliverable

Decision Analytica is a web application that you will run on your PC's browser.

The application will be responsive i.e it can be used on PC or mobile but specifically designed to be used on PCs. The reason is that we will be dealing with large set of alternatives and criteria for user's problem. The large matrices will be formed showing different comparisons. It seems inappropriate to use it on mobile. As large matrices may collapse.

Decision Analytica is being developed using PHP Laravel as a front-end and back-end framework. We are using Visual Studio Code to write a code. Some APIs for implementing algorithms will be developed in Django using Python.

The front-end is being developed with the latest bootstrap 4, HTML5, CSS on Laravel version 8. The version for PHP is 7. MySQL will be used as a backend database.

It's an interactive application where user will be provided with a different questionnaire to give their opinions. To implement these questionnaires we will be hosting a separate website for the questionnaire form which will display questions and then answers will be saved in the database. The project needs to be hosted on one site. We need to buy amazon hosting for that.

The Deployment Diagram shows how different parts are interacting with each other. Web browser is shifting control to Application layer. Here our business logic is implemented. Then the data after processing goes to Database system.

We will be providing a complete commited website after passing it through different Test Phases.

Final Deliverable of the Project Software SystemCore Industry ITOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Decent Work and Economic Growth, Industry, Innovation and Infrastructure, Reduced Inequality, Sustainable Cities and Communities, Responsible Consumption and ProductionRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	80000
Printing	Miscellaneous	20	500	10000
Amazon Web Hosting	Equipment	1	30000	30000
Cloud Hosting Services	Equipment	3	10000	30000
Subscriptions	Equipment	1	10000	10000