Agent based model of effects of task allocation strategies in flat organizations

• An agent based model of a stylized, call-center-like organization is presented.
• The common practice of piling work on best performers is found to be effective.
• An equal distribution of tasks among agents is less effective.
• Power laws are found in transition times to an overload under variable global load.

A common practice in many organizations is to pile the work on the best performers. It is easy to implement by the management and, despite the apparent injustice, appears to be working in many situations. In our work we present a simple agent based model, constructed to simulate this practice and to analyze conditions under which the overall efficiency of the organization (for example measured by the backlog of unresolved issues) breaks down, due to the cumulative effect of the individual overloads. The model confirms that the strategy mentioned above is, indeed, rational: it leads to better global results than an alternative one, using equal workload distribution among all workers. The presented analyses focus on the behavior of the organizations close to the limit of the maximum total throughput and provide results for the growth of the unprocessed backlog in several situations, as well as suggestions related to avoiding such buildup.