Online published on 28 February, 2020.
This thesis deals with two of the most important competency mapping aspects of faculty members in technical institutes: Allocation of academic load to faculties using computational intelligence and subsequently, determining variable pay of faculties in proportion to performance. In this study both the objectives have been dealt with proven computational approaches. The motivation behind the objectives stemmed froma veryappalling problem i.e.lackof employability skills in technical graduates. The root cause of the problem reveals several contributing factors, but the one factor which can be isolated, is the impact of the faculty members. The faculty members are at the helm of an eûective teaching learning process and have a major impact on students’ achievements. This necessitates that the faculty members be motivated enough to have the required impact. In this context two aspects that are of paramount importance for the performance of a faculty are: being in the right place in terms of teaching load and getting rewards in tandem with performance. The research was conducted using various dataof facultymembers ofatechnicalinstitute. Inthe wake of the two aspectsmentioned above a Multi-objective particle swarm optimization model was proposed to optimally allocate academic load considering several objectives like maximizing the preferences of individual faculty members, depth of knowledge of the teacher in that subject, contribution to the area of the subject through research or publication and minimizing the variance in load of diûerent faculty members. The objectives were also associated with several constraints. The objective of the second model was to compute the variable pay component of the faculty members using Sugeno fuzzy inference system.
The aim was to overcome the pitfalls of the forced distribution system in place and make a fair computation of the variable pay in tandemwiththe deûned performance parameters. The performance parameters were based on the Academic Performance Indicators (API) laid by UGC. The problems might seem trivial at the outset but the amount of data to be processed, number of objectives and constraints, rules and the precision required in the decisions increase the complexity of the problems. Hence the proposed models have been based on sound and proven computational approaches which avoid errors related to biases in human judgment and limitation of human mind in processing large amounts of data and rules simultaneously. Finally both the models are sustainable in the sense that they can be used repeatedly with or without further customization. The simulation results reveal that the proposed models fair well in terms of the objectives. The academic load allocation model yields solutions which satisfy the required objectives of maximizing the preferences of individual faculty members, depth of knowledge of the teacher in that subject, contribution to the area of the subject through research or publication and minimizing the variance in load of diûerent faculty members and handle the constraints very well. The fuzzy inference model resulted in 48% of the faculty members logging increase in the amount of variable component in comparison to the traditional allotment and 10% of the faculty members who were not getting any variable pay amount got some amount as per the proposed model. The proposed models can be used by any academic institution with little or no customization. The models were implemented in MATLAB and ûndings of the study were analysed in MATLAB and MS Excel. Though two very important facets of the problem discussed at the outset have been dealt with, it is imperative to further this study, so that a computational intelligence (CI) based competency framework can be built for institutions. Hence the gaps and several research directions based on these gaps have been proposed which satisfy the third objective of this research i.e. identify research gaps which can be pursued for developing a CI based frame-work for academic institutions.
