A Machine learning model for task allocation

Abstract

Task allocation is one among the key planning exercises that plays a major role in an organization’s quest to satisfy Service Level Agreements and to attain operational excellence. Task allocation is a difficult issue that addresses the inter-dynamics of tasks and employees, having in mind factors such as skills utilization, fairness and diversity. Conflicts arise when tasks take too long to be resolved due to in expertise and poor task allocation. The individuals assigned end up not performing the task well and incomplete tasks/project are presented. The end user is usually dissatisfied and will end up giving negative feedback. Previous studies have not addressed the issue of matching two or more employees to the task with qualities possessed by the employee, preference of the team or partner given and the qualifications for the task given. The nature of this problem has within this research been equated to the stable marriage problem solved by the Gale and Shapley’s Algorithm. However, a serious concern of the Gale-Shapley algorithm is its non-truthfulness also known as the man optimal result. Gale-Shapley’s algorithm was used to bring out the human aspect during task allocation. The aim of this research is to formulate an algorithm to support task allocation problem using machine learning. To ensure time efficiency, a predictive machine learning algorithm (artificial neural network) has been used to show improved time with each pair towards a task. A mixed method research methodology was used whereby it combined elements of quantitative and qualitative research approaches. The precision of the developed model was at 0.713421 bringing the estimate of the regression model to 71.34%. The findings showed that the neural network can be used as an effective algorithm for predicting the task allocation.