25 th Australasian Conference on Information Systems Effort Estimation for Production Support Planner 8 th -10 th Dec 2014, Auckland, New Zealand Jha et al. An Agent Based Approach for Effort Estimation in Production Support Ashish Kumar Jha System Research Laboratory Tata Research Development and Design Center Pune, India Email: ak.j@tcs.com Abhinay Puvvala System Research Laboratory Tata Research Development and Design Center Pune, India Email: abhinay.puvvala@tcs.com Veerendra K Rai System Research Laboratory Tata Research Development and Design Center Pune, India Email: veerendrak.rai@tcs.com Sanjit Mehta System Research Laboratory Tata Research Development and Design Center Pune, India Email: sanjit.mehta@tcs.com Harrick M Vin System Research Laboratory Tata Research Development and Design Center Pune, India Email: harrick.vin@tcs.com Abstract Estimating the effort required in resolving an incident for a production support engagement has been a long standing problem. Although, a lot of work has been done in software engineering field in estimating effort required for software development, there is no adequate body of work on production support effort estimation. Reliable effort estimates for handling each kind of incidents can have the benefits of planning the complete project engagement for many services firm which thrive on production support projects. This paper presents a framework for estimating mean effort required to resolve an incident. It takes agent based modelling approach to capture the attributes and interactions of the agents. The framework has been tested with few policy experiments to assess the amount of effort required to resolve the incidents and its impact on SLA compliance. The model has been tested with industry data and holds promise to be a reliable tool to conduct policy experiments and enhance resource utilization. Keywords Software Maintenance, Production Support, Incident management, Effort Estimation, Optimization, Agent Based Modelling INTRODUCTION With increasing automation and thrust for use of software, software projects are rarely small and never simple. A lot of inherent complexities are getting built in owing to its increasing ubiquity. Maintenance of such software encompasses handling incidents, which is an event that is not part of standard operation of a service and causes or may cause an interruption to or reduction in quality of that service” (Kapella, 2003 pp.2). Such a comprehensive software maintenance contract is often called as production support or production management. Average maintenance cost of software over its complete life cycle can be as much as 90% of the total cost (Wiederhold, 2006). In such a scenario the importance of production support becomes immense and has to be