Developing Cement Supply Simulation Model to Optimize Scheduling and Vessel Capacity Utilization

Developing Cement Supply Simulation Model to Optimize Scheduling and Vessel Capacity Utilization


Cementos Argos is a cement and ready-mix market leader in Colombia and the fourth largest cement producer in the United States. It has over 8,000 employees and $3 billion in revenue.

It operates across 16 countries and includes 13 cement plants, nine grinding facilities, 340 concrete plants, and 33 ports and terminals. Argos provides cement for both the industrial and retail segments. The specialists create value for the customers through innovative products and solutions.


Argos has two bulk carriers and 31 terminals. Each of the terminals has different silo storage capacity, draft and port restrictions, and terminal operations that have been considered in the model.

Argos engineers would model cement supply in Excel, using updated reports for each terminal: Daily Sales, Inventory Report, and Inventory Planning. Moreover, they did the updating and planning every week, or on demand.

Considering the system constraints, the specialists would manually build up a supply plan. These constraints were known by experience rather than modeled in any system. Modeling in Excel gave the visibility of only 15 days from the operations planning to execution.

To improve the visibility and the accuracy of the overall cement supply planning process, Argos opted for simulation modeling in AnyLogic.

For the simulation modeling project, Cementos Argos had three main objectives:


Using AnyLogic, Argos developed and analyzed a simulation model for the supply of bulk cement to the Eastern Caribbean terminals. The model was then exported as a standalone Java application.

AnyLogic agent-based simulation model

AnyLogic agent-based simulation model (click to enlarge)

Argos’ bulk carriers moved cement from a plant in Cartagena to terminals in Aruba, Dominica, Antigua, St. Maarten, and St. Thomas. These locations were mapped with the GIS Map shape in the simulation model.

On the model settings screen, users could enter the information about terminals, including variables such as initial inventories and projected demand in each market segment. The model also considered such parameters as storage capacity, packing rate, and working hours.

For vessels, model input data contained the following variables: assigned routes, proposed volume split, speed ranges, and departure delays. It also considered vessel parameters including: capacity, and loading and unloading rates.

Alternatively, the simulation model could read the input data from, and export it to, the same Excel file.

The developers used AnyLogic Fluid Library to simulate vessel loading and unloading as well as cement storing, packing, and dispatching from the terminals. They also used AnyLogic schedule elements to simulate the terminals’ working hours.

The terminals’ schedule

The terminals’ schedule (click to enlarge)

The Cementos Argos engineers used an agent-based simulation approach and statecharts to model the vessels’ behavior. To simulate stochastic parameters (vessel speed, cement demand, etc.) they applied built-in probability or custom distributions . Users could also specify the simulation starting date and simulation time (in hours).

The statecharts of the vessels’ behavior

The statecharts of the vessels’ behavior (click to enlarge)

The engineers could view the simulation results on charts and in tables.

For the cement terminal, the model outputs included:

For the vessel, the output data contained:

 Model outputs for vessels and terminals

Model outputs for vessels and terminals (click to enlarge)


Simulating cement supply operations, Argos extended their visibility from 15 to 30-35 days. The company also created and analyzed different cement supply scenarios using AnyLogic. The model developers could easily test the scenarios, make necessary adjustments, and eventually choose the best option.

Thanks to simulation modeling, vessel monitoring and sales and production updates were carried out monthly, rather than weekly, which saved a lot of engineers’ time.

To measure the model’s performance, Argos engineers defined KPIs and compared the results to the objectives.

Argos KPIs

Argos KPIs

The results with the simulation KPIs:

Benefits of the simulation model

Using AnyLogic, Argos optimized its cement supply scheduling and vessel capacity utilization. Also, simulation helped to review the system bottlenecks and identify possible risks such as inventory shortages, shipment delays, and so on. This improved product availability in the real system.

Moreover, modeling reduced labor costs and overtime. As a result, in 2021, Argos sold higher volumes than in 2020 with the same unit cost.

Next steps

Argos engineers were going to add autonomous destination definition for vessels in the model. This way the vessels could choose optimal next destination themselves.

The engineers would also like to combine heuristics with the reinforcement learning technology to increase vessel (agent) choice efficiency.

Watch the video about this case study presented by Cementos Argos at the AnyLogic Conference 2021.