Decision-Making in Partially Observable Environments

Organizations are striving to find solutions that perform their business processes more efficiently and more effective. While organizations pursue these objectives, the surrounding and the internal environment is changing and consequently incrementing the complexity for the managers to formulate full informed decisions. Moreover, the organizational environments do not allow to full recognize the overall surrounds or the internal current operation, e.g., large fusion and/or acquisitions, or complex semi-automatic business processes. This paper combines partially observable Markov decision processes with DEMO business transactions to support the management initiatives with more knowledge about the impact of their actions in the operation of the organization. An agro-food company shows the usefulness of the solution and the results forecast the management decisions value in situations where is not possible to fully observe the state of the reality, or to be fully aware about it.