Traditionally, assembly plants have been characterized as being rigid structures,where a product type always follows the same production path. This structurelimits the order of operations and machines that a product can follow, which couldrestrict the room for improvement in reducing the overall production time. The needfor this flexibility becomes of paramount importance when one or more machines arecapable of performing the operations that were supposed to take place in a brokendown machine.This thesis studies one specific solution to this problem, the substitution of therigid production lines with a modular flexible assembly system based on stationsdistributed along the plant. This distribution requires the use of Automated GuidedVehicles (AGVs) to transport the products from one station to the next one througha transport network.This structure allows not only to increase the range of different products that can beoffered due to the flexibility of the system, but also, if controlled correctly, minimizesthe production delays when a machine failure occurs, redirecting the production toother machines. The increase in flexibility causes an increase in the complexityof production management and control. A whole new set of variables comes intoplay. AGVs transport times, AGVs availability, products availability, deadlocks,etc. are some of the variables that need to be taken into account when optimizingthe schedule. This complex system makes imperative the development of algorithmsable to schedule, coordinate and control the production in an optimal way.In this context, the purpose of this thesis is to develop two algorithms capable ofoptimizing the production schedule and able to react to unexpected events duringproduction control. A performance analysis to compare both algorithms will bedeveloped as well under the same conditions and in different scenarios using arealistic simulation developed for this purpose. These algorithms are based on twodifferent structures: the dispatching algorithm using a decentralized structure anddispatching rules to decide which product should go to which machine, and theImperialist Competitive Algorithm based on a centralized structure and the ICA asan optimization algorithm, which is an evolutionary algorithm.The results show a better performance of the ICA algorithm in all tests. This is dueto the need for a 100% deadlock avoidance system and a better AGV managementin terms of predicting where a future transport will be needed