Let's keep doing what we’ve been doing and focus on doing it even more effectively. We have built up momentum for implementation of operational simulations in the right direction and substantially improved performance. If external factors don’t disrupt our progress, the right thing to do is remain focused and keep moving in direction of proven results.
Here, we provide summary of the areas in which momentum has been achieved in creation and implementation of new simulation models and what efforts must be sustained. Most areas of progress have their origin in Site Visit Executive initiatives.
At any given time, up to one-third of the destroyers, cruisers, and amphibious ships are under significant repair.
Some have party tents erected on their decks, others are encased in scaffolding, and a few are lifted completely out of the water on massive drydocks — all signals that overhauls are under way. Ships can remain like this, unable to deploy, anywhere from weeks to years, depending on the scale of the maintenance.
When you have assets like that tied up, it’s hard on the Navy.
So Navy is looking for ways to shorten the repair time for not just ships, but planes and combat vehicles too.
The ideal behind having less time in maintenance is that in effect you have a larger Navy because there’s more ships at sea. The whole thing is around speed. How do we get speed?
About one-quarter of the Navy’s surface ships are currently going through extended maintenance periods that last anywhere from six months to a year. During that period, major components, like engines, are overhauled. Even ships that aren’t in this extended period of downtime undergo about three to four weeks of maintenance quarterly while in port.
Getting these ships, planes, and combat vehicles out of overhaul faster frees them up for training and deployments, thus boosting readiness and lethality.
The instability — in terms of the availability of ships and scheduling — is probably one of the more complicated aspects of this. If you could get something that’s smooth, in terms of backlog and schedule for the suppliers and contractors, they’re going to be a lot more productive.
Aircraft are a different story. It’s been widely reported how maintenance issues have grounded nearly two-thirds of the Navy’s strike fighters. The military will use an new, computer-based database to manage logistics, maintenance and the supply chain of its new F-35 Joint Strike Fighter. But there have been numerous problems bringing the system online.
Instead of being sub-optimised, how do we want to run some of these areas like an Business because the availability is greater. What are the things that we can steal from industry best practices that lend themselves to solving some of these systemic problems?
Shipbuilding is a little bit different than aircraft building, than is different than ground vehicles, but there are a lot of these practices that lend themselves to the other business.
A key to making it all work is having the budget to pay for the maintenance work. If it takes less time to go through maintenance, it costs less. If it costs less, there’s more ships available and the shipyard can put more ships through there, everyone is a winner.
Covering all requirements of modern military industrial enterprise simulations is not easy problem. We propose two agent-based technologies for manufacturing support on two different levels: intra-enterprise and extra-enterprise level-- can be used simultaneously or together.
The standard configuration consists of several independent systems linked to the virtual organisation by agent-based decision support technology in various fields of manufacturing – maintenance planning, supply chain decision-making models, simulation, extra-enterprise access, etc.
Integration of the Simulation exercise that we have carried out at different industries validated that the agent-based technology is viable in situations where the maintenance planning problem is constantly changing e.g. project driven production, and requires frequent and continuous re-planning.
In these Readiness Simulations the collective aspects of the agent technology have been exploited. At the same time we have identified a great potential of the technology in situations where the planning problem is characterised by complex processes but it features some of the internal logistics.
This has been case of the high volume production/maintenance availabilities, where not only collective aspects of agent simulation technology have been used but also the integrative capabilities of agents have been exploited e.g. integration of the linear programming heavy-duty solver.
In simulations where the planning metrics are widely distributed and not fully available the agent technologies provide an robust integrative and distributed planning framework for supply chain administration and virtual production/maintenance organisations formation.
The multi-agent approach provides a specific modeling and simulation alternative to the known mathematical and system science modeling technologies for simulating the manufacturing process.
Here we describe single Phases of cooperation life-cycle on Enterprise-to-Enterprise level searching of the possible product support collaborators. First, agents have to contact possible partners. There is wide field for future research in the domain of automatic searching and contacting possible partners.
This approach ensures the trustworthiness of the partners transferred from real-life to the agents cooperation. Each agent is equipped by the addresses and the security certificates and every partner can be authenticated using standard key methods. Every agent can be connected to many partner agents according to defined internal cooperation rules.
Once the agents are connected together, each agent provides the list of available product support capabilities to partners. It is possible to propose different capabilities to different partners. During this phase agents form basic cooperation network, receiving information suitable for effective collaboration in the next phases. During the life-cycle of the cooperation, agents subscribe information of the changes on product support resources on already established cooperation.
Agent that discovers a need for outsourcing of a part of their activities starts looking for the best possible partner for the cooperation. First, the agent searches its local product support network for all cooperators, which are potentially able to agree on collaboration.
Secondly it negotiates with selected partners about details of a possible collaboration. Ones the cooperation is agreed by both sides, virtual organisation is established and product support contract become standard operating procedures for all involved partners.
The originator is responsible for using of and paying for agreed product support upgrade capacities and the cooperator is responsible for providing it. Even if the conditions are changed by one of the partners, agents tries to keep the contract.
Simulation provides for production/maintenance re-planning once the cooperation is settled, with agents informing each other about every relevant change. If the initiator requires a change of product support contract conditions, it informs the subcontractor about its requirements and subcontractor tries to meet the new specification.
If the subcontractor can finish its sub-task sooner or later then agreed, it immediately informs the task originator. When one of agents goes off-line, the connection is delayed and during a next successful connection all accumulated changes are exchanged.
Any partner as well as some kind of independent product support organisation can run agent feedback to monitor and evaluate any cooperation, like asking for communication logs, which can or may not be provided. Available product support metrics can be used for evaluation, measurement and future optimisation of cooperation.
The heart of our product support administrative actions is set of planning agents using manufacturing case-specific approaches ie, decomposition based planners or heavy duty planners, and set of resource agents.
Here we describe roles of individual agents:
Multi-Agent Robotic Systems Information exchange in real world design scenarios can be used for simulation and modeling of production process, product support and associated supply chain, where they easily simulate an independence of critical parts involved in the operation.
These tools can help to answer non-trivial tasks – how changes in single component will affect the production process or product support supply chain as a whole.
Task sequencing is key to achieve task prioritisation because it affects the order in which equipment is allocated and used, and potentially which pieces of equipment are available at each point during mission execution.
If tasks occur sequentially, equipment used in one task may be available in the next. However, if tasks overlap, then equipment required by multiple tasks may only be available for one activity, forcing substitution and reallocation.
Site Visit Executive is able to maximise readiness and overcome equipment shortfalls by manipulating the timing and sequencing of tasks/subtasks involved in operational availability scenarios. Can involve reordering certain tasks over others or staggering tasks rather than attempting to execute them concurrently.
The framework design uses heterogeneous system, to enable connecting new units like robotic agents. It sets a decentralised network for communications between agents, avoiding the need for permanent communications.
Operational records are created based on the last information transfer and the sensor indicator. To achieve this, it is important to describe the required characteristics for agents in along with a description of framework processes.
In Multi-Agent Robotic Systems, information distribution and processing are autonomous due to their modularity and distributed architecture. System modularity allows the system to be robust because it can detect and easily replace agents or parts of them that are not working. Also, if the system needs to be partially updated, only the necessary agents need to be changed, reducing work efforts.
The Point-to-Point transmission model does not require constant communication with the whole system, which does not need a central unit reducing the bottlenecks in the communication system, which is ideal for Multi-Agent Robotic Systems. This model must be complemented by a communication protocol for information transfer.
Here is decentralised framework for product support tasks that enables the connection of heterogeneous agents to the results in ability to perform multiple tasks simultaneously without depending on a global control. The approach to communication between agents is through a Sensor Network since they are comprised of nodes with independent processing units, wireless communication modules, and sensors.
We used this technology expand the applications of Multi-Agent Robotic Systems. Sensor networks are used to receive information from multiple agents to increase resource use and efﬁciency of equipment tasked in the administration of product production/maintenance processes.
In this particular case, robotic agents are static actuators that, although sharing information locally, depend on a single control system for information processing without requiring a permanent connection to work. This is possible through the use of non-infrastructure networks that do not depend on a central unit for info processing.
The framework presented here is a procedure for decentralised communication between agents for purpose of product support monitoring.. Modular process framework can be modiﬁed at any moment on a system or level agent. However, when the system is ﬁrst started, modular processes must be executed sequentially.
The ﬁrst process is the characterisation of the agents components and their relationship with product support job site criteria; the second is assessments of transmissions between agents; the third is initialising agents for their connection to administrative network; the fourth process is the generation of the product support activity history for each agent in the system. Finally, the ﬁfth process describes information transmission through a header to be included in the information packet.
Marines can readily utilise simulation of dispatching information packets to be sent downrange to complete discharge of a weapon system without needing to validate their correct reception.
Higher transference rates are allowed for, but at the same time Marines run the risk of increasing error rates and preventing information loss by validating each simulation packet sent between agents. However, in cases where agents have a weak link, this is reﬂected in a higher packet loss rate, increasing time when there is an undeﬁned conﬁrmation loop between agents.
Marines are building short courses to teach many ranks how to best use training simulations to create realistic war fighting exercises.
For far too long there has been a disconnect between the gaming capabilities of Marines entering the Corps and how the war fighting simulators are used to build their training.
But a group at the Marine Corps Air Ground Combat Center is changing that dynamic with short-form courses that create limited experts at the squad and battalion level.
In effort to test the how well simulation application can support mission planning, we created scenarios designed to stress its ability to assess equipment shortfalls.
Instead of simply executing a single textbook operation, we explored the increasingly prominent split–Amphibious Readiness Group operating concept, as well as the more complex challenge of multiple simultaneous operations.
We used traditional combined readiness group as our control group, and examined individual missions in order to be able to compare their impact on equipment availability with multiple simultaneous missions.
Until now one of the only ways for a commander to test his Marines via simulation ahead of a field exercise or during downtime would be to seek the limited resources of officers and staff at a place like the Battle Simulation Center at the Marine Air Ground Task Force – Training Command.
That’s fine for large-scale work. But we must bring the capability down to individual platoons and squads to help Marines operate in the field or a real-world event.
The Corps continues to use simulators that have been in service for more than a decade and also add items like Tactical Decision Kits, a combination of drones, cameras and laptops that allows Marines to scan an area ahead of a mission and do dry runs virtually.
Mainres have shrunk the major themes down to a course called “Simulation Professional Course” and a shorter course called “Simulations Specialist Course” to gives Marines a deeper dive into simulations and is focused on creating a battalion-level expert who can translate commanders needs into a simulated training package.
The idea behind the course is to have an enabler who understands what it takes to put together training objectives, understand simulations and put training together for any level of staff or Marines.
This isn’t just us playing video games in a Sim Center somewhere. Using the simulations helped Marines better understand their place in a combined arms exercise, bringing such training to individual squads and translating commander goals into simulated exercises.
Our goal is to expand training sessions and spread that number across the force until there are enough Marines trained in the art of simulation to make a difference in success rates of critical mission sets.
1. Considerable overlap in the types of tasks and activities involved in the mission set.
2. Commonalities important to mission planning because imply similarities in equipment requirements may also exist
3. Application allows equipment allocation to be constrained, facilitating planning under suboptimal conditions
4. Application allows allocations specify operational conditions effect on equipment requirements
5. If tasks occur sequentially, equipment used in one task sequence could be available in the next
6. If tasks overlap equipment required by multiple tasks may only be available for one activity forcing substitution/reallocation
7. Mission tasks may exhibit relative priority so some tasks may be more important than others.
8. Prioritising tasks ensure most effective pieces of equipment are available to complete most important tasks
9. Some scenarios may require unit to complete more than single mission from operation set
10. Complex operations involve several overlapping missions to be completed sequential/simultaneous