Abstract
In this paper, we are interested in the modeling and the resolution of the dynamic and discrete berth allocation problem which is noted DDBAP. To resolve this problem, we propose a heuristic approach of optimization which combines two concepts: agent and heuristics. This approach is based on the use of the multi-agent negotiation, the contract net protocol, and a set of heuristics such as the WorstFit arrangement technique and the LPT policy. The objective of our work is then, to solve the problem of scheduling n tasks on m parallel identical machines. The criterion that we aim to minimize is the makespan (in analogy with the P parallel to Cmax problem) having a set of constraints to be satisfied. We developed our model of negotiation using the Jade platform. We finish this work by presenting various simulations to show the performance of the proposed heuristic and the contribution of our approach compared to other already existing approaches.