In this paper we introduce an enhanced methodology to determine the optimal control strategy for a complex trigeneration plant. The plant is designed to meet the thermal and electrical loads of a user and is connected to the electrical grid. We consider a single working-day and the plant set-points are determined on an hourly basis minimizing total energy cost, plant maintenance costs, and costs associated to switching on and off the power plant components. To realistically simulate the behavior of large power plants, a constraint on the minimum duration of on and off intervals is considered for each plant section. The problem in study is discretized in time and plant states, represented as weighted graph, and the strategy that minimizes the total cost is determined using backward dynamic programming, whose computational effort is compatible with real practical applications. Validity and usefulness of the proposed methodology are demonstrated optimizing the set-point of a combined heat, power and cooling system, under different seasonal load conditions and energy prices. We demonstrate that an optimized strategy would reduce the total daily cost from 8% to about 100%, depending on seasonal load, with respect to rule based control strategies, such as heat-tracking and electrical tracking. Crown Copyright (C) 2014 Published by Elsevier Ltd. All rights reserved.