Self learning approach in software design

Today the software engineering is in a very advanced stage. The various concepts of software engineering made the software development process very simplified. But most of the software engineering concepts are focused on the developing the programmed systems, which suffer limitations in comparison with the human experts. If we want to approach near to human decision making and learning ability, we need to design a self learning system. Self learning systems can save lot of system development and system upgrade time. To design a self learning system we have very good reference model: the human brain. We know that the human brain has very complex structure. On the other hand we also know that it has the most simplified algorithms to perform complex tasks. We do not need to focus on the complex structure of the brain to design the self learning approach, but on how the brain uses these simplified algorithms for intelligent decision making. During our research we proved that the brain uses mathematical optimisation similar to mathematical techniques we use very often. Once things and actions are represented mathematically they can be easily transferred into any self learning system. Our experiments proved that actions, decisions, behaviours, human’s nature, logic, feelings and thinking are the result of mathematical calculations on the knowledge base acquired by each individual. Human decisions are based on calculations supported by knowledge base and case specific database of an individual. The knowledge base is generated through the experience, learning and training. We know that even the simplest decision making involves lots of conditions and parameters. Evaluating all the parameters and conditions is very much tedious task without the mathematical method/tool. The software like Oracle Business Objects, SAP, already proved that the decision activity of the brain can be transferred to the system, but all these systems evolved keeping the “Human Assistant” approach in mind. We have not yet started relying on this kind of systems, but given a chance these system can be improved to the self decision making level. One of our successful attempts is that we have drafted the digital decision making architecture of brain. It follows similar approach to computer and modularity in software development. The only difference is that software works on programmed systems and it implies the high level of accuracy, human brain works with approximation which implies less accuracy but a very wide scope for learning and adaption. With a self learning and adaptation architecture it is possible to link a traditional programmed systems with the proposed software design. This approach is applicable every area of software enabled entities