PoDMan: Policy Deviation Management

Abstract

In an organizational context, policies are a collection of high-level, operation and management goals and/or rules which are used to define the expected behaviours and processes of organizational actors. Hence policies are concerned with ordinary, non-exceptional business processes and how organizational actors perform these processes. However, in every organization, unexpected events do occur from time to time which require deviation from established policies, and currently there is very little known in the literature on how the actors can proceed under such situations. This thesis aims to propose a method which includes a multi-step framework to design and develop applications that can assist organizational actors in deviating from policies while facing unexpected or exceptional situations by allowing collaborative policy-based information systems to identify policy deviations by its users, present them with alternate actions and behaviors to attain the required outcomes and, when necessary, allow them to deviate from the policy and perform a non-conforming action. This study has been conducted by following Hevner’s framework for design science research which suggests obtaining awareness of the problem from the environment, putting forward suggestions to solve the problem, contributing to the pool of study through publications and evaluating the solution to validate the generalizability and correctness of the solution.

The outcomes from this research are, a policy deviation framework based on existing literature and a multi-step policy deviation method which can assist organizational actors in detecting and deviating from policies when necessary. The methods proposed in this study have been evaluated using hypothetical scenarios based on real policies and through randomly generated simulations, and has been shown to be both applicable and effective.

In terms of contributions to research, the fields of CSCW, software agents, decision making, as well as policy based planning and management can benefit significantly from this work.