1. Introduction

The defense M&S is considered to be attractive in the sense that it may simulate war operations virtually in advance, and it may lead to the cost saving and human being safety for the actual operations and for the preparative training as well. It includes all activities of physical, mathematical, logical modeling and simulation in defense area in order to achieve the military training effect without actual field training. Sustainability requires organizational reusability base for problem- solving, which is usually built after extensive trial and error learning. It is also emphasized to facilitate implementation of interoperability and reusability strategies in order to promote sustainability in defense modeling and simulation (M&S, hereafter). The defense M&S has been heavily studied and discussed both in academia and in practice such as in the reusability and interoperability of defense area [3, 7, 24, 28, 29].

A defense modeling is a simplified representation of actual military operations to promote understanding of military operations [2]. A defense simulation is the manipulation of a modeling enabling one to perceive the interactions of components in models to replace the actual military field training without serious problem due to the availability of many advantageous techniques [19]. Also, complicated systems such as software utilized by weapon system should be reused on other projects whenever technically possible. However, this objective is not easy to achieve practically in most cases. To achieve this goal, system should be developed from scratch for the purpose of the system reuse. In addition, the basic principle of interoperability should be followed carefully to other project for the purpose of successful reuse of system. The conceptual model of the mission space (CMMS, hereafter) has to be designed and developed so that it may secure the interoperability and reusability, which are the most important characteristics the conceptual modeling [18].

This paper suggests the relevance and practicality of CMMS-K (Conceptual Model of the Mission Space-Korea), a basis for a common methodological framework for the preliminary steps to get a clearer understanding of the concept and war simulation model development for the mission space. It is designed to meet the general requirements for the simulation modeling and the special constraints of the Korean mission space, in consideration of the particular circumstance of potential war tension. The CMMS acts as a bridge between the real world of the military and the simulation developers. It is very important to defense M&S interoperation and reuse. A military domain scenario is action-centric [18, 16]. In order to maximize the reusability of CMMS-K, this paper proposes the task ontology which describes an action and a task. The task ontology defines a task as a group of actions which are performed by the same subject repeatedly [1, 5, 20, 22, 26]. This approach applied in implementing a normalized scale military M&S environment by reusing existing agent-based simulators. The proposed approach is useful for modeling complex systems and facilitating interoperation between the model of conceptual level and between simulators of implementation level

2. Related Work

2.1 History of CMMS

Due to the complexity and disparity of the technology in defense M&S area, architecture is becoming increasingly important. Acronym CMMS has been replaced by other ini-tiatives several times. FDMS RC (Functional Descriptions of the Mission Space Resource Center) had suggested the new terminology FDMS for CMMS, which was integrated with MSRR (Modeling and Simulation Resource Repository) to have a model of KIRC (Knowledge Integration Resource Center) about 2002. CMMS, FDMS and KIRC have all been introduced by DMSO (Defense Modeling and Simulation Office, US) [31].

In 1999, the ADSO (Australian Defense Simulation Office) decided to establish the draft simulation master plan and a defense strategic plan. CMMS for M&S requirements is described and redefined with a standardized description for entities, operational environments and tactics for each service needs. ADSO has tried to provide the concept of CMMS for Australia in Defense M&S master plan. The Swedish defense research agency (FOI) has accepted the idea of the CMMS in simulation model for the defense area because of the recognition of the interoperability and reusability. As a result, the FOI carried out a study since 2001, to investigate what extent and under which circumstances the concept could be of interest for the Swedish defense. The result of the study led to the development of DCMF (Defense Conceptual Modeling Framework). The NATO (North Atlantic Treaty Organization) RTO (Research and Technology Organization) MSG (Modeling and Simula-tion Group) recognized the conceptual modeling in interoperable simulation assets, and thus set up the NMSG CM (NATO Modeling & Simulation Group Conceptual Model) - 058 in 2008 [9, 10, 12, 13].

2.2 CMMS at DMSO (U.S.)

The primary and the most important part of CMMS includes: domain specific conceptual models, so called Mission Space Models (MSM). The other parts are the Technical Framework, a common library with a database management system, and a group of supporting tools for different parts of the CMMS. For the implementation of standardization, the MSM structure is composed of Conceptual Model (CM), Data Interchange Format (DIF) and Data Base Management System (DBMS).

CM is composed of a Common Syntax and Semantic (CSS) for representing all activities related to the military operations. DIF interchange CSS's contents to mechanical representation as standard format of the mechanical switch technology. DBMS serves as a data repository and the external environment is recognized as a component. CM can be represented as a standard framework through the structure of the CSS. The detail concept of a CSS is using the terms relationships. The term of CSS is composed of the entity, state, verb, conditions, actions, actor, processes, tasks, and use cases. Interrelationships of all terms represent the military environment of mission space and each entity can perform the role depending on the value of the property [8, 14, 21].

While the conceptual models of CMMS and DCMF can describe objects, actions, and processes [6, 27], they do not support an efficient mechanism for reusable and interoperable actions [4, 30]. They have a limitation that they cannot store and search sets of actions which are used repeatedly. BOM++ of FOI is a representative model description lan-guage and extends the conventional Base Object Model (BOM). BOM is a Simulation Interoperability Standards Organization (SISO) standard and encapsulates information needed to describe a simulation component. BOM is defined in the form of XML (Extensi-ble Markup Language). To describe a CMMS, BOM++ proposes two alternative approaches to extend BOM. In the first one, existing implementations remain undisturbed and the schema structure of BOM is extended and refers to an external ontology [11, 23, 32, 33]. In this paper, we have proposed an efficient conceptual modeling method of describing a CMMS-K based on the proposed task ontology and the control constructs of OWL-S.

3. Conceptual Models of the Mission Space - Korea (CMMS-K)

Interoperability and reusability of models are main concerns in military modeling & simulation area. High interoperability quality in defense M&S environment is a technologically complex task, being affected by multiple factors, and the task is not yet satisfactorily characterized. In order to improve the interoperability and reusability of models, the model shall be conceptualized with a common concept, and the framework of models as well as the architecture of the simulation system should be standardized. Interoperability in defense M&S area is considered a critical capability for future joint forces. This approach proposes Interoperability reference models for the exclusive use of military applications, taking into account the specific requirements of military environments.

Conceptual Models of the Mission Space-Korea (CMMS-K) has its origin in the CMMS concept presented by US DMSO. It constitutes an important step in the implementation of the Korea Defense's modeling and simulation plan by initiating the first study of how a common library of verified and validated conceptual models of military operations can be developed. ADD (Agency for Defense Development, Korea) has started research on the concept to explore its potential. ADD began with an extended study of known published material about CMMS. Conceptual modeling was discovered and it was revealed that many specifications of CMMS were vague and some part has still a lot of room for improvement. In this section, CMMS-K as suggested provides more practical framework of conceptual modeling and more advanced task ontology.

3.1 Framework of CMMS-K

Conceptual Model (CM), Data Interchange Formats (DIF), and Data Base Management System (DBMS) are configured in CMMS-K. In framework of CMMS-K, CM is expressed with concrete objects, task and process through interaction between them for the mission space. CM and relevant details for each component are as follows:

The basic component of object is composed of an entity, expressed by target state, verb, and condition. Entity includes human, organizational troop, equipment, facility, etc. which are representative of the entity. State represents the attribute and property of internal and external environment of an entity, indicating its status. For example, the terrain location of a tank is downtown. Verb means the subject of an action, represented by movement of the entity, such as, departure, detect, communicate, reconnaissance, etc. Entity condition is a particular behavior, for example, when tank is in open fire condition, it could tell whether to shoot in standstill state or in driving state. Description section shows the content of chosen items for each user's pleasure during final action.

The task is the smallest and basic unit of realization of the mission space, expressed by relationship of several object, each object has a meaningful action to achieve the military target of mission space. Standardization and generalization concept of task list could be more efficient to achieve the mission space with an access to authorized military knowledge [17, 25]. CMMS at DMSO had not enough concept of task list. UJTL (Universal Joint Task List), standardization and generalization list by national development levels, is not appropriate intact for being applied to CMMS because of its ambiguity and the expression of mission space can be restricted by necessary details.

3.2 Tactical Task Ontolgy

The CMMS-K domain ontology consists of entity (object and location) and task ontologies. In this section, the tactical task ontology is proposed in military M&S areas. The Korean tactical task list consists of maneuver, information, firepower, combat service support, defense and command / control.

The main components of the tactical (TA) task ontology can be classified into classes and properties. These classes are classified again into actions and tasks. An action is a minimum unit of activity which constitutes a task. A task is a group of actions conducted by the same subject to fulfill a mission. The properties are classified into action properties and task properties for representing the task ontology of CMMS-K. The main action properties are described in <Table 1>.

Actions and tasks have a hierarchical structure. This means that a child task (action) has a 'is-a' relationship with its parent task (action). A task (action) is more general than its descendant task (action). For example, task Target-Acquisition has its child tasks: Tar-get-Acquisition-With-Human Information and Tar-get-Acquisition-With-Detection-Equipment. Especially, a task with its sibling tasks which have the same parent task should include at least one action that is not included in any of its sibling tasks. The reason is that defining a task separately from its sibling tasks is meaningful only when the task has an action that is not included in any of its sibling tasks. For instance, task Target-Acquisition-With-Detection- Equipment has action Detection-Equipment-Inspection that is not included in its only sibling task Target-Acquisition- With-Human-Information. Likewise, a child task should include at least one action that its parent task does not have. This makes a more specific definition of a child task than its parent task. A child task Target-Acquisition-With-Detection Equipment includes action Detection-Equipment-Installation and its parent task Target-Acquisition does not have the action.

All actions that are included in a parent task are inherited to all of its child tasks. The reason is that a parent task is a more general concept than its child tasks. For example, if task Target-Acquisition has action Target-Decision, all of its child tasks, Target-Acquisition-With-Detection- Equipment and Target-Acquisition-With-Human-Information, inherit the action. Even if a property range of a task (action) is not specified, its child task (child) may have a described range of the same property. On the other hand, if a property of a task (action) has a range of an entity, its child task (action) should have the same entity or a child entity of the entity as a range of the property. It is because a task (action) is a specialization of its parent task (action). The information described by the main task properties of the tactical task ontology cannot be expressed with previous conceptual models in the military domain such as BOM++ pro-posed by FOI. Additionally, the tactical task ontology provides grammar that can express more information, which is meaningful, than previous military conceptual models. For instance, action property has-Indirect-Object, takes-Place-In-Location and has-Effect of the tactical task ontology describe an indirect object, location and effect of an action respectively which cannot be represented with BOM++ grammar.

<Table 2> represents task Target-Detection-And-CASRequest as an example to illus-trate the task ontology. <Table 3> describes action CAS–Request-Decision which constitutes task Target-Detection-And-CAS-Request. Close Air Support (CAS) is air action by fixed-wing and rotary-wing aircraft against hostile targets that are in close proximity to friendly forces, and requires detailed integration of each air mission with the fire and movement of those forces. Task Target-Detection-And-CAS-Request has unit as a subject.

The task consists of action Target-Detection, CAS-Request- Decision and CAS-Request. The ultimate purpose of the task is fire. Duration of performing the task is not described in class level while duration can be described in duration type in instance level. This paper deals with description in only class level. A precondition is also not described in class level. In the case of successful completion of the task, an effect is described in rule language which connects two statements 'Unit has target information.' and 'CAS request is passed to squadron.' with connector 'and'. The ranges of task property has-Precondition and has-Effect of the tactical task ontology are described in rule language in the tactical task ontology but are described in natural language in this paper for easy understanding. In the case of unsuccessful completion of the task, no particular effect of the task is described in class level. In addition, this task depends on none of the other task.

<Table 3> represents the property ranges of action CAS-Request-Decision. The activity of this action is verb ‘decide’ and the subject of this action is unit. This action takes another action CAS-Request as a direct object and does not take any indirect object. No specific resource is used for a subject to perform this action in class level and location to perform this action is also not determined in class level. This action takes target information as input. The precondition of the action is ‘Unit has target information.’ The action produces output CAS request decision information in the case of successful completion of performing the action and output CAS request decision failure information in other case. If the action is successfully completed, the following effect is ‘Unit decided to request CAS.’ If not, no specific effect is described in class level.

4. Mission Space Modeling: A Case Study

This section concentrates on describing a CMMS-K based on the proposed task ontology. Section 4.1 describes an example CAS scenario as case study. Section 4.2 represents the example CAS scenario as conceptual model using OWL-S based on the task ontology. The results demonstrated the potential of ontology-based systems to promote interopera-bility, exchanging information and pointing out the inconsistency within a defense M&S system.

4.1 Close Air Support(CAS) Scenario

This paper used an example CAS scenario based on military publication, while CAS scenarios can be described in various forms in a real world. The CAS (Close Air Support) engagement situation is supposed as an engagement scenario. The concept of the scenario is shown on <Table 4>.

<Table 5> shows the simplified task ontology of the example CAS scenario. Tasks and actions in <Table 5> are extracted from the example CAS scenario in <Table 4>. 1~9 of the example CAS scenario in <Table 4> correspond to action 1~9 in <Table 5>. In this scenario, action Aircraft-Scrambling- Command and CAS-Scrambling-Command relate to connector choice and all the other actions relate to connector sequence.

4.2 Conceptual Model Using OWL-S

Conceptual model uses OWL-S control constructs to describe models based on the task ontology. A conceptual model diagram representing the example CAS scenario based on the simplified task ontology in <Table 5>. Using parameter bindings of OWL-S, the followings could be described. In composite processes, the input to one process component can be obtained as one of the outputs of a preceding step, or the outputs of a composite process may be derived from outputs of some of its components.

As shown in <Table 6>, a scenario can be described using parameter bindings of OWL-S which shows data flow. For example, output target information of action 1 becomes the input of action 2. In addition, output CAS request information of action 3 is used as input of action 4~9. Using this parameter binding grammar of OWL-S, data flow of this scenario could be easily identified. To describe the example CAS scenario, our team modified OWL-S as follows. As parameter bindings could exist in composite OWL-S processes, state bindings also might exist between the precondition of an action and one of the effects of its preceding action. If parameter binding exists between Effect1 of Action1 and Precon-dition2 of Action2, and if Effect1 is false after performing Action1, it can be inferred that Action2 cannot be carried out because Precondition2 is also false. For example, the precondition of action CAS-Request is 'Unit decided to request CAS.' which is the effect of ac-tion CAS-Request-Decision. As a result, the precondition and the effect are connected with state binding, and if the effect is false after performing action CAS-Request- Decision, it can be drawn that action CAS-Request cannot be performed because its precondition is not satisfied.

Users easily can use components as plug-in and play way because they are made of dynamic link library in military area. As the dynamic link library is a binary code, the logic and source code of model can be also protected from leakage. By the result of this approach, CMMS-K can give the basic simulation environment for the enhancement of reusability and interoperability of models, especially in Korean military applications. In additionally, the developed CMMS-K system is a tool to transform conceptual level model into implementation level models, making them more reusable and more interoperable in the context of larger multiperspective in defense M&S area.

5. Discussion and Future Research

The application possibility of CMMS-K for interoperability and reusability in Korean Army is identified. Without these efforts of standard, it is almost impossible to develop to reusable entities and a model of components due to their feature of embedding various core framework or simulation system service within the code for coordinating their activity. Interoperability between entities and components is impossible due to the fact that each simulation system providing its own event processing system bearing with its own unique interfaces. Thus, it can be estimated that defense M&S must recognize that models are different from general software components and that model composability needs to be based not simply on software practice, but on the converging science of defense M&S area.

The suggested framework and task ontology of CMMS-K is being developed to reflect the characteristics of military environments while referring to existing military conceptual modeling frameworks. CMMS-K can reduce the cost of actual military trainings in terms of time, space and supplies. For objective validity and practical application, it can be con-firmed by applying this concept through the user interface as developed after applying the appropriate practices for long term national security. CMMS-K can be presented for the actual implementation and handling of the interoperability and reusability issue in defense M&S area. For more practical applications, the task ontology technology as a way to implement the CMMS-K ontology based on OWL-S is presented with example scenario. The task ontology describes an action and a task, a group of actions, systematically and exquisitely which are absent in previous researches. Using the task ontology, a group of actions with the same object which are repeatedly used together could be described as a task, storing and searching tasks is possible. It can be identified as a potential extension in CMMS-K management system of military applications including search, delete, operate, reuse, editing and save by the military environment change. Reusability of the knowledge produced at the end of this approach can bring numerous benefits to the military M&S area. Some ontologies are available for semantic content management of CMMS-K knowledge but in order to avail their full benefits, it is must that they can dynamically link the library in various military M&S systems.

It is increasingly common for defense M&S planners to consider the entire range of doctrine, technic, environment, and economic means for achieving military goals so as to enable what is called reusability-based planning and interoperability- based operations. Thus, there is a need for M&S capabilities that can address this broader context in a coordinated and sensible manner while taking into account the learning and adaptation that take place a route to attaining the desired reusability and interoperability. Applying the CMMS-K, concept and its verification are still a potential area of future research. Other ways for improvement of interoperability and reusability in defense M&S also exist but CMMS concept is very successful when applied to advanced cases of defense M&S area. Here as elsewhere, it is possible to check indirectly the practical applicability of the CMMS-K. Taking it a step further, the suggested CMMS-K has the capability to act as a more practical method during actual level application. Additionally, it could help making the simulation software easier to develop, use, reuse and maintain, and achieve both a higher quality and higher level of interoperability and reusability at a reduced cost in defense M&S development process.