YushiLan1,,Kebo Deng,*,Shaojie Mao,Heng Wang,Kan Yi,and Ming Lei

1.The 28th Research Institute,China Electronics Technology Group Corporation,Nanjing 210007,China; 2.Science and Technology on Information Systems Engineering Laboratory,Nanjing 210007,China

Adaptive evolvementofinformation age C4ISR structure

YushiLan1,2,Kebo Deng2,*,Shaojie Mao2,Heng Wang2,Kan Yi2,and Ming Lei2

1.The 28th Research Institute,China Electronics Technology Group Corporation,Nanjing 210007,China; 2.Science and Technology on Information Systems Engineering Laboratory,Nanjing 210007,China

Command,control,communication,computing,intelligence,surveillance and reconnaissance(C4ISR)in information age is a complex system whose structure always changes actively or passively during the warfare.Therefore,it is important to optimize the structure,especially in ambiguous and quick-tempo modern warfare.This paper proposes an adaptive evolvement mechanism for the C4ISR structure to survive the changeable warfare.Firstly,the information age C4ISR structure is de fi ned and modeled based on the complex network theory.Secondly, taking the observe,orient,decide and act(OODA)model into consideration,four kinds of loops in the C4ISR structure are proposed and their coef fi cient of networked effects(CNE)is further de fi ned.Then,the adaptive evolvement mechanisms of the four kinds of loops are presented respectively.Finally,taking the joint air-defense C4ISR as an example,simulation experiments are implemented,which validate the evolvement mechanism and show thatthe information age C4ISR structure has some characteristics ofsmall-world network and scale-free network.

C4ISR structure,complex network,loop,adaptive evolvement,coef fi cient ofnetworked effects(CNE).

1.Introduction

Command,control,communication,computing,intelligence,surveillance and reconnaissance(C4ISR)is an integrated military information system for battle fi eld information collection,procession,dissemination and utilization. With the developmentof information technology and military requirement,especially the proposition of networkcentric warfare(NCW),C4ISR has gradually transformed to a network-centric system of the information age from a stovepipe system of the industry age[1-4].Meanwhile, the C4ISR structure(communication network,information interaction relationship,software invocation relationship, etc.)has also developed from hierarchical,preplanned and fi xed to fl at,dynamic con fi gurable and adaptive[5-7]. Compared with the system of the industry age,the information age C4ISR system enables its structure to evolve dynamically during the combat,that is,adjust the unit function orthe relationship between the units as the task or battle fi eld situation changes[8-11].Structure evolvement is importantto make the bestof information age C4ISR in ambiguous and quick-tempo modern warfare,such as deploying mobile radar temporarily in hot area to enhance the target detection ability,establishing new intelligence supportrelationship to meet the commander’s unexpected information requirement,recon fi guring the command relationship if some command postis destroyed.

Present researches on the C4ISR structure mainly focus on the evaluation of the static structure.Dekker applied socialnetwork analysis concepts to the C4ISR structure[12,13],and proposed the force,intelligence,networking and command and control(FINC)methodology, with which different C4ISR structures could be evaluated and compared from the facets of delay analysis,centrality analysis and intelligence analysis.The FINC model has been well used in static C4ISR structure evaluation [14-17].Scheidt and Schultz used the information theory to investigate the utility of alternative command and control(C2)structures,such as centralized control,scale-free hierarchies,scale-invarianthierarchies and“small world”[18,19],and found thatthe C2 structure should vary as the situation changes,leaving the method of varying not further studied.Nowadays,more and more attention has been paid to the method of constructing an agile C2 structure to fi t uncertain and dynamic scenarios in the future[20,21]. For example,in recent years,the International Command and Control Research and Technology Symposium(ICCRTS)has taken the“agile C2 structure”as the main topic, including the methods of building and evaluating the agile C2 structure[22-25].However,how to optimize an integrated C4ISR structure during the warfare has notyetbeen studied thoroughly.

This paper studies an adaptive evolvement mechanism of the C4ISR structure in order to make the best use of the information age C4ISR.In Section 2,the informationage C4ISR structure is de fi ned and modeled based on the complex network theory,with four kinds of units and four kinds of relationships between them being covered.Taking the observe,orient,decide and act(OODA)[26,27] model into consideration,Section 3 proposes four kinds of loops in the C4ISR structure,including collaborative observation loop,collaborative decision loop,command and control loop and collaborative action loop,and further de fi nes their coef fi cient of networked effects(CNE). Then,the adaptive evolvement mechanisms of four kinds of loops are presented respectively in Section 4,including the facets of evolvementstimulus,goal,condition,course and rule.Taking the joint air-defense C4ISR as an example,Section 5 discusses the simulation experiments implemented,whose results show that the structure evolvement mechanism could improve the C4ISR ef fi ciency.Finally,it concludes that the C4ISR structure has some characteristics of small-world network and scale-free network during the evolvement.

2.Structure ofinformation age C4ISR

Information age C4ISRis builtbased on the military information infrastructure.The military information infrastructure(e.g.,global information grid)can be considered as a wide-area interconnection and resource-sharing environmentcomposed ofthe facilities ofcommunication,computation,storage,information service,management,security, and so on[28,29],as the bottom layer shown in Fig.1. With the basic military information infrastructure,upper application units(information source,command post, weapon equipment,etc.)can con fi gure a variety of logic relationships,such as intelligence support,situation awareness sharing,command and control,and collaborative action,as the middle layer shown in Fig.1.These logic relationships among the upper application units can also be adjusted according to the task and the environment,as the middle layerto the upper layer shown in Fig.1.This paper only focuses on the upper application units and their logic relationships,and de fi nes them as the C4ISR structure, excluding the bottom military information infrastructure. Therefore,the C4ISR structure here is composed of C4ISR units and their logic relationships,described as follows.

Fig.1 Structure of information age C4ISR

2.1 C4ISR unit

The C4ISR unitis physically independent,with some speci fi c function and certain responsibility.It can be classi fi ed into four categories:battle fi eld-observation unit, intelligence-processing unit,decision unitand action unit. (i)The battle fi eld-observation unitcollects the attribute information including shape,position,electromagnetic spectrum parameters and so on of military targets in the battle fi eld via sensors.Radar,sonar,reconnaissance satellite and electronic supportmeasure(ESM)allfallinto this category.

(ii)The intelligence-processing unit integrates intelligence from multiple battle fi eld observation units,and gen-erates real-time,continuous,clear and accurate intelligence oftargets.Differentclasses ofradarintelligence processing systems,electronic countermeasures intelligence processing systems,technical reconnaissance intelligence processing systems and intelligence processing centers all fallinto this category.

(iii)The decision unitanalyzes and understands the battle fi eld intelligence sent from the battle fi eld observation unit or the intelligence processing unit,makes the operation scheme,and commands the subordinate force to act. Differentclasses of command posts in differentarmed services allfallinto this category.

(iv)The action unit follows the order or scheme from the decision unitto accomplish the operationaltasks,such as differentkinds oftroops,combataircraftand airdefense fi repower units.

2.2 Relationship

Relationship refers to the interaction between C4ISRunits. There are four basic kinds of relationships within the information age C4ISR structure:intelligence support relationship,command and control relationship,collaborative relationship,and backup relationship.

(i)The intelligence support relationship is the relationship between the intelligence provider and user,which is established according to task responsibility.To be speci fi c, the units who possess intelligence resources continuously send necessary intelligence associated with the task to the units who need it via the network.For example,there is an intelligence support relationship between the radar intelligence processing center and the aviation division command post.

(ii)The command and control relationship exists between the decision unit and its subordinate force,or different classes of decision units,which is achieved by the combatorder,plan and state feedback.For example,there is a command and control relationship between the aviation division command postand the combataircraft.

(iii)The collaborative relationship refers to the coordination and mutualassistance during the operation between two units withoutthe superior-subordinate relationship by sharing situation awareness and operation plan.For example,there is a collaborative relationship between two neighboring aviation division command posts as they try to interceptthe intrusive aircraftin the common area.

(iv)The backup relationship is the relationship between one unitand its backup unit.For example,an aviation division command post will replace its neighboring destroyed command post.Similarly,if an intelligence processing unit fails,another processing unitwillalso replace it.

Based on above fourkinds ofunits(O,P,D,A)and their relationships(R),this paper models the information age C4ISR structure(denoted as OPDAR)by applying the theory ofcomplex network,as shown in Fig.2(b).Further,the mathematicaldescription ofthe structure-adjacency matrix can be derived from this model,as shown in Fig.2(c).

Fig.2 Expressions of information age C4ISR structure

3.Loops in C4ISR structure and CNE

The loop in the C4ISR structure is a closed path composed of the units and their relationships.There are four basic types of loops in C4ISR structure:collaborative observation loop,collaborative decision loop,command and control loop,and collaborative action loop.For example,a tactical level air defense C4ISR has four radar stations(O1,O2,O3,O4),a radar intelligence processing center(P),an aviation command post(D2),a jointair defense command post(D1)and two interception aircraft (A1,A2).The radar stations,intelligence process center and command posts are interconnected via the military information infrastructure,while the command post and interception aircraft are interconnected via data link.The four basic kinds of loops are de fi ned as follows.

(i)The collaborative observation loop,which is composed of multiple networked observation units and intelligence processing units(Fig.3(a)),provides collaborativeobservation of networked observation units and the feedback control to the observation units according to the intelligence processing result.For the collaborative observation loop,there could be collaboration among heterogeneous or homogeneous sensing assets.For example,the radar detects the target collaboratively with the ESM,the early-warning aircraftindicates the targetto the fi re control radar,and several early-warning air defense radars jointly detectthe intruded aircraft.Besides,itis also possible and importantto optimally allocate and controlthe sensing assets according to the battle fi eld observation result.All of these can enhance the target detection ability and quality, enabling information superiority.

(ii)The collaborative decision loop,which is composed of multiple decision units(Fig.3(b)),allows situation awareness sharing and collaborative decision between different decision units,enabling decision superiority by common operationalpicture(COP)with high quality,situation awareness,decision and combatplan.

(iii)The command and controlloop,consisting of multiple decision units and action units(Fig.3(c)),serves two major functions:i)the superior decision units command and control its subordinate decision units or action units; ii)the subordinate gives feedback of their states and warfare circumstances.Itenables real-time monitoring of battle fi eld and timely adjusting of the execution.

(iv)The collaborative action loop,which is composed of multiple networked action units(Fig.3(d)),conducts collaboration among differentaction units in the same task to synchronize operation.The command and controlloop and the collaborative action loop directly affectthe acquiring of action superiority.

Fig.3 Basic loops in the structure of a tacticallevelair defense C4ISR

From the previous analysis,itcan be concluded thatthe four basic kinds of loops can ef fi ciently enhance the abilities of C4ISR from various aspects.The more loops,the more information exchanging,situation awareness sharing,feedback control,collaborative decision and action in the C4ISR.Therefore,C4ISR is more enabled by the units networking.Shannon drew the same conclusion in the research of the distributed networked operation[30]. Since the loop number in the C4ISR structure can be measured approximately by the maximum eigenvalue(denoted asλmax)of the adjacency matrix(if there is no loop in the structure,λmax=0;if only one loop exists in the structure,λmax=1;if multiple loops exist in the structure, λmax＞1),this paper usesλmaxto de fi ne the CNE of the C4ISR structure.As a result,the comprehensive capability (denoted by Yca)of C4ISR can be expressed as

In the above equation,the constantαdenotes the loop’s impact on C4ISR ef fi ciency.The larger the value ofαis, the more important role the loop plays.n denotes the total unit number in the C4ISR system,and f(Xa1,Xa2, ...)denotes the overallcapability of C4ISRwithoutloops.Adaptive evolvement of the C4ISR structure can be concisely expressed as:to improve the comprehensive capability of C4ISR,quickly and automatically adjust the unit function and relation in different loops as the task or situation changes.The following section will elaborate the evolvementmechanism of differenttypes of loops respectively.

4.Adaptive evolvementprocess and mechanism of C4ISR structure

Generally,there are four task stages for the C4ISR system to complete an operation,including intelligence collection and processing,decision support,command and control (assigning operational plans and instructions),and action. These stages are linked,supported and fed back by each other according to the OODA-loop model.In these stages, the missions of the C4ISR system are relatively independent;therefore,if the adaptive evolvement of the C4ISR structure is studied within an operation,it can be divided into four sub-stages on the timeline.The structure evolvementmodelateach task stage includes evolvementdrivers, optimization objectives,constraints and structural adjustmentmodels.

Taking air-defense C4ISR as an example,this paper introduces the dynamic evolvement model of the system structure at each task stage.The initial system structure consists of organic units.Assume the system has one joint air-defense command post(D1)with three subordinate defense areas,each ofwhich hasfourradarstations,one radar intelligence processing center,one aviation command post and severaloperationalaircraft(A).Fig.4(a)shows the initial C4ISR structure of a certain defense area when facing a mission,with T standing for targets,O1-O4standing forfour radar stations respectively,P1standing for the radar intelligence processing center,and D2standing for the aviation command postin this defense area.According to the battle fi eld situation,the joint air-defense command postcan manage allradar stations,intelligence processing centers and aviation command posts of its three defense areas to participate in the air-defense operations.

4.1 Evolvement characteristics of collaborative observation loop

At the stage of intelligence collection and processing,the task objective of C4ISR is to organize different kinds of observation resources to obtain the best battle fi eld intelligence.In essence,it is the collaborative observation loop which evolves as the situation changes by optimally organizing and applying the sensing resources.Therefore,the subjectofthe structure evolvementis the collaborative observation loop atthis task stage.

The driving factors of the dynamic evolvement of the collaborative observation loop lie in that the target observation quality cannot meet the requirement of current mission,or some observation units are disabled or degraded due to electromagnetic interference or physical attack.Constrained by the communication capability of the military information infrastructures and the effectiveness of the observation nodes,the evolvementcan be achieved by means of adding new intelligence collection units apart from the organic units,or enhancing the collaboration and management of the intelligence collection units.The evolvementofthe collaborative observation loop of the airdefense C4ISR is shown in Fig.4,in which O5and O6stand forthe radarstations subordinated to neighboring defense areas.

Fig.4 Evolvement of collaboration observation loop

In Fig.4,itis obvious thatadding new intelligence collection units can increase the CNE of the system structure. However,since it is only the overallevolvementprinciple and there are always multiple units thatcould be chosen,a mathematicaloptimization modelis required to selectnew intelligence collection units.This section will propose amodel to solve this problem.The optimization objectives take the observation quality of the targets,the cost of the observation and the types of working nodes into consideration,while the constraints include the transmission delay of intelligence reported by the observation nodes,the effectiveness of the observation nodes,etc.

Let G0={O1,O2,...,On}represent the initial set of the organic observation units included in the collaborative observation loop,G＇=stand for the available set of the observation nodes apart from the organic units,and{M1,M2,...,Md}be the set of targets. The optimization objectives and constraints are explored as below.

(i)Optimization objectives of the cooperative observation loop

i)The observation quality

Since the quality criteria such as integrity,accuracy and clarity of the fused battle fi eld image depend on the target detection probability,the jointdetection probability of the targets can be chosen as the optimization factor of the observation quality.Suppose a new set of observation nodes G+is added into the initialobservation loop G0and forms a new collaborative observation loop G0+={G0,G+}. Pi(Oj)denotes the detection probability of the target Miusing observation unit Oj,then the joint detection probability of the target Miwith a new collaborative observation loop willbe

Since the roles of different targets and their expected detection quality are different,the requirements of the observation quality of different targets can be distinguished by assigning weights to them.Assuming that the target Mihas the weight Wi,the normalized observation quality can be expressed as

Obviously,in the evolvementprocess of the collaborative observation loop,the ultimate goalis to maximize the overallobservation quality,denoted by max P(G0+).

ii)The observation cost

To choose the observation nodes,the observation quality,the possible interference/attacks caused by the exposure of the radiant source,as well as the extra time for intelligence processing caused by the increased data from sensors should be taken into account.Therefore,while pursuing the observation quality,the number of observation nodes as wellas theirworking time should be decreased to reduce the observation cost.Assume thatthe costofobservation node Ojis C(Oj),whose value depends on the time increment(Δtj)of the intelligence processing caused by adding the node Oj,the possibility of interference/attacks (pj)caused by the exposure,and the possible consequence (cj)due to the interference/attacks.Then C(Oj)can be expressed as C(Oj)=f(Δtj,pj,cj),where f stands for an increasing function with a positive range.Then the observation cost of the new collaborative observation loop (G0+)can be normalized as

where N denotes the observation unit number in the new collaborative observation loop.

Obviously,in the process of dynamic evolvementof the collaborative observation loop,the ultimate goal is to minimize the observation cost,denoted by min C(G0+).

iii)The types of the observation units

There are various types of observation units such as radar,photoelectric observation and passive observation, which can collect target features in various aspects.For example,radar can precisely measure the location and motion characteristics(distance,position,height,velocity,acceleration,etc.)of the target,butcan be threatened by the jamming,low-altitude penetration,and anti-radiation missile.Infrared sensors,with good covertness,high angular resolution and strong anti-interference performance,can measure the infrared characteristics of the targetunder the condition of poor visibility.ESM can measure the parameters of a radiant target,including amplitude,frequency, pulse width,pulse repetition frequency and azimuth ofsignals,and further identify the targetwith these parameters. Let the total type number of the observation units in the new collaborative observation loop G0+be|G0+|,this optimization factor can be normalized as

where G＇indicates the setof observation units outside the observation loop.

From the perspectives of the precision of target location,the accuracy of attribute recognition,the credibility of threat evaluation and the anti-interference capability,we intend to maximize the number of observation unit types in the collaborative observation loop,denoted by max L(G0+).

(ii)Constraints of the collaborative observation loop

i)The time effectiveness of data transmission

Based on the military information infrastructure,the observation loop will aggregate the observation information and send it to the intelligence processing unit.Since various observation units share the same communication resources atthe same time,the information fl ow willincrease as the number of observation units increases,thus intensifying the competition of communication resources and fi nally in fl uencing the time effectiveness of data transmission.The time effectiveness is essentialfor the observation loop,especially for time-sensitive targets.The increase of transmission delay will decrease the value of the target’s information.Extremely,when the transmission delay exceeds a certain threshold,the observation intelligence will be useless.Suppose the allowed information delay threshold is T0,the information delay of the observation unit OjisΔt(Oj),then the constraint can be described by

ii)The availability of the observation unit

At least one target must exist in the observation range of each collaborative observation unit,otherwise the node will have no contribution to the collaboration,thus can be ignored.Assume thatthe observation range of node Ojis ρj,the distance between Ojand the target Miis dij,and then the constraint of the availability of the observation node can be described by

where f stands for

iii)The capacity of intelligence processing

Adding new observation unit will notonly increase the transmission load ofthe communication resources,butalso intensify the workload of the intelligence processing unit. Since the capacity of the intelligence processing node is limited,information broughtby the newly added observation node may exceed its processing capacity,thus delaying the processing of inputinformation.Therefore,the input capacity of the intelligence processing unit should be considered as a constraintwhen optimizing the observation loop.Assume that the processing capacity is Cpro,which could be the number of targettracks or the information inputintensity,and the amountof data from observation unit Ojis hinput(Oj),then the constraintcan be expressed as

(iii)Optimization modelofthe collaborative observation loop

According to the previous analysis,the optimization factors of dynamic evolvement of the collaborative observation loop include observation quality,observation costand the types of observation units.Among them,higher observation quality,lower observation cost and more types of observation nodes willcontribute to the optimization.The constraints include the time effectiveness of data transmission,the capacity of intelligence processing and the availability of observation units.The time delay of data transmission mustbe shorter than expected,the amountof data processing mustbe smallerthan expected,and the observation units mustbe available.Thus,the optimization model can be de fi ned as

In the evolvement of the observation loop,appropriate observation nodes will be chosen based on the optimization modelabove.This modelcan be solved by the genetic algorithm(GA),colony algorithm or simulated annealing algorithm.

4.2 Evolvement characteristics of collaborative decision loop

At the stage of decision support,the task objective of the C4ISRsystem is to generate a correct,clear,integrated and consistentcommon operationalpicture(COP)or common tactical picture(CTP)in time,assisting commanders in making the best decisions.This objective can be achieved by the timely and broad sharing of situation awareness and collaborative decision making among the decision units based on the localbattle fi eld postures.Therefore,the subjectof the evolvementin this stage is the collaborative decision loop.

The evolvementdriving factors of the collaborative decision loop lie in jointoperation,ambiguous battle fi eld situation,and precon fi gured intelligence supportfailure.The evolvementaims atobtaining and understanding the battlefi eld COP/CTP quickly and correctly by situation awareness sharing.The constraints of the evolvement include the communication capacity ofthe military information infrastructure,the command system and institution,and the operational organization.The methods of the evolvementinclude interaction with new decision units,extending the situation awareness range and adjusting collaborative decision relations.Fig.5 shows the collaborative decision loop of the air-defense C4ISR system,with D3and D4representing aviation command posts af fi liated to neighboring defense area organic units.

Fig.5 Evolvement ofcollaboration decision loop

Fig.5 indicates that interacting with new decision units can improve the CNE of the system structure.However, this is only the overall evolving principle of the decision loop,choosing new decision unit is constrained by command relationship and tasks.If many units are available, the optimization model should be applied.It is expected thatthe newly added decision unitwould contribute to the quality of battle fi eld situation awareness.And this contribution can be mathematically modeled by the entropy theory[31].Battle fi eld situation awareness can be regarded as the recognition/identi fi cation of allrelated targets in the battle fi eld,including their types,positions,status,capabilities,and intentions.Assume that Mijstands for the attribute j of target Mi,and the local judgment of decision unit Dkon Mijis p(Mij=q|Dk),i.e.Mijhas the value q with the possibility of p.Based on the Dempster-Shafer(DS)evidence theory,the overallunderstanding of n collaborative decision units for Mijcan be deduced,denoted by p(Mij=q|D1⊕D2⊕···⊕Dn).Then,the information entropy of the variable Mijcan be computed, denoted by h(Mij=q|D1⊕D2⊕···⊕Dn),and the information entropy of all concerned targets,i.e.the battle fi eld circumstance can further be computed,denoted by h(Ω|D1⊕D2⊕···⊕Dn).The formulas are

When there are n decision units available for adding, the one making the greatest contribution to the battle fi eld situation awareness can be determined by the formula

4.3 Evolvement characteristics of C2 loop

Atthe stage of C2,C4ISR intends to continuously perform ef fi cientcommand and controlby quickly issuing accurate combat plans and instructions to the forces or the combat platform and monitoring the operation process in order to synchronize the operation among the involved forces.As a result,the evolvement core at this stage is the C2 loop whose evolvementis driven by two aspects.First,according to the mission,itrequiresto command new forces.Second,due to some decision units’failure and ineffectiveness,the command relationship should be reorganized in orderto maintain ongoing command.

In modern warfare,especially in joint operation,thecommand post should continuously reorganize the forces under control.Within command capacity,increasing the types and numbers of commanding objects enlarges the commanding range,and thus enhances the combatability. In terms of command structure,simplifying the levels and shortening the period lead to better command abilities and higher ef fi ciency.During the operation,the evolvementof the C2 loop is driven by combatplans or instructions,and thus the action units cannot join the C2 loop until receiving combatorders.Fig.6 depicts the evolvementof the C2 loop in tactical levelair defense C4ISR,with A3,A4,A5, A6representing operationalaircraftaf fi liated to neighboring defense areas respectively.

Fig.6 Evolvement of C2 loop

During the operation,the C2 loop should be evolved to maintain ongoing command when the conventionalcommand chain is broken by some decision units’ineffectiveness due to failure or attack.According to the levelof the invalid decision unit,differentevolving patterns ofbypassing command,upgraded command and transferred command as shown in Fig.7 are available for application.In details,if the decision unit of the top levelfails,upgraded command is the only choice.Either bypassing command or transferred command can be adopted when the decision unitofthe bottom levelfails.Ifthe decision unitin the middle levelfails,three patterns are all available.With regard to upgraded command and transferred command,several decision units are available for taking over command,in this case,the optimization modelis required according to the intelligence processing ability(Cip),information interaction ability(Cii),remaining C2 capacity(Ccc)and signi fi cance(Rp).Suppose K units are available,the optimization can be modeled as

whereλidenotes the weight,represent normalized intelligence processing ability,information interaction ability,remaining C2 capacity and signi fi cance of k respectively.The calculation formula is as follows:

Fig.7 Evolving patterns of C2 loop with some decision units’failures

4.4 Evolvementcharacteristics ofcollaborative action loop

Atthe action stage,C4ISRstructure evolvementmainly focuses on the collaborative action loop.C4ISR in this stage intends to monitor the actions of operational force and supportthe collaborative operations among action units to achieve synchronous action.From the perspective of joint operations,close cooperation between the action units will bring aboutadvantages of jointoperation,greater than the sums of each action unit.With more action units joining and closercollaboration between them,the CNE ofthe collaborative action loop willincrease,and the capability and ef fi ciency to accomplish an operational mission will thus be improved.Therefore,at the action stage,the evolvementprocess of the collaborative action loop can be taken as constantly adding new action unitand establishing new collaboration relationship among the action units to form a new collaborative action loop with higher CNE.During the operation,the evolvementof the collaborative action loop is driven by the operational plan and the real-time tacticalsituation,and constrained by the communication capability of the battle fi eld communication assets and weapon platforms,such as the maximum node number supported by the data link and the ability of data transmission.Fig.8 represents the evolvementof the collaborative action loop.

Fig.8 Evolvement ofcollaborative action loop

4.5 Essentialconclusion of C4ISR structure evolvement

The core of the C4ISR structure is a cluster of nodes and links,which plays a key role in improving system ef fi ciency.According to previous sections,the structure evolvementcore exists at differentpositions in the C4ISR structure at different stages.At the stage of intelligence collection and processing,the core of the system structure is collaborative observation loops;at the stage of decision support,the core is collaborative decision loops;at the stage of operation command,the core is C2 loops;at the stage of action,the core is collaborative action loops.

The basic principle of the adaptive evolvement of the C4ISR structure can be summarized as follows.During an operation,as the requirement of the task or circumstance changes at different operation stages,different loops quickly and automatically adjust their unit joining and relationship between the units in an environment restricted by resource capacity,task requirement and command mechanism,in order to optimize the C4ISR structure with comprehensive capability improvement.The adaptive evolvementof information age C4ISR is shown in Fig.8. Fig.9(a)shows an initialsystem structure designed for an air defense task,in which the CNE is 2.39.Fig.9(b)is the evolvementresultatthe stage ofintelligence collection and processing,where the dotted rectangle denotes the system structure core at this stage.The CNE of the system structure increases to 2.72.Fig.9(c)indicates the structure evolvement result at the stage of decision support,where the structure core has moved from the collaborative observation loop to the collaborative decision loop,and the CNE increases to 3.02.Fig.9(d)represents the evolvementresultatthe stage of operation command,where the core has moved to the collaborative C2 loop and the CNE increases to 3.35.Fig.9(e)shows the evolvement result of system structure at the stage of the action,where the core is the collaborative action loop and the CNE increases to 3.73.

Fig.9 Adaptive evolvement process of C4ISR structure

5.Simulation experiments

Taking area joint air defense C4ISR as the simulation object,the experiment evaluates the dynamic evolution features of the system structure by constructing and operating the joint air defense simulation system.This experiment consists of three parts:(i)justifying the core shifts phenomenon of the system structure during operationalapplication;(ii)verifying the advantages broughtto the system ef fi ciency by the dynamic evolution of the system structure;and(iii)analyzing the features of classical complex networks presented by the C4ISR structure during its dy-namic evolution.

The simulation experiment sets basic parameters of 30 intelligence collection units(radar,ρj∈[250 km,500 km],Cpro＞100),15 intelligence processing units(radar processing center,intelligence fusing center),10 decision units(joint air defense command post, aviation command post,and ground-to-air missile command post)and 200 action units(ground-to-air missile, aircraft).Among these,some units are not part of the organic units,and do not participate in the combat mission at the initial stage.System units mentioned above are deployed on the same military information infrastructure,enabling information sharing and collaborative operation among them to supportthe dynamic evolution of the system structure.Some simulation parameters are set as α=1,Wi=1,T0=2 s,N=30,m=60.In the scenario,60 threattargets(Mi,i=1,2,...,60)are approaching the air defense area atthe speed of 800 km/h as shown in Fig.10,and the expected value of the target detecting coverage coef fi cientis setas 3

Fig.10 Scenario ofsimulation experiment for system structure dynamic evolvement

5.1 Adaptive evolvement of loops and core shifts of system structure

According to Section 3 and Section 4,the core of the C4ISR structure plays a key role in enhancing networking ef fi ciency(the maximum eigenvalue of 0-1 adjacency matrix of the system structure)from the perspective of complex network;and it also contributes to operational ef fi ciency from the perspective of operationalapplication. For convenience,this experiment explores the core from the perspective of operation in terms of the fl ow and fl uctuation of information transmitted among different kinds of loops,in order to verify the phenomenon of core shifts during the evolution of the C4ISR structure from both aspects mentioned above.

During the joint air defense simulation experiment,the C4ISR structure keeps evolving as the warfare proceeds, resulting in the changing of CNE of different loops.As shown in Fig.11,CNE of differentloops are allenhanced during the evolvement,and fi nally reach a stable condition,which means differentloops are optimized by evolution.On the other hand,differentloops are optimized one by one,and the core of the structure at the initial stage is collaborative observation loops,then collaborative decision loops,command and controlloops,and collaborative action loops.However,the optimization process of different loops may be overlapped.The experiment concludes thatthe evolution of the C4ISR structure is optimization of differentloops and continuous shifting of the core.

Fig.11 Changing of CNE of different loops over simulation time

In an analysis of core shifts of the system structure in terms of information fl ow and its changes,the information types include assurance intelligence,situation awareness, instruction and order,and collaboration,corresponding to differentkinds of loops in the system structure.Assurance intelligence mainly corresponds to the collaborative observation loop composed of intelligence collection units and intelligence processing units;situation awareness mainly corresponds to the collaborative decision loop composed of decision units;instruction and order mainly correspond to the command and controlloop made up ofdecision units and action units.In the experiment,we collect the message sentand received by each unit,gatherstatistics of exchanging fl ow among different types of messages periodically and analyze the changes of message interaction fl ow oversimulation time.As the results shown in Fig.12, at the initial stage of the simulation experiment,the information fl ow of the system mainly occurs in the message of assurance intelligence whose fl ow per unittime increases signi fi cantly;indicating that at this stage the collaborative observation loop is the core of the system structure.When simulation time exceeds 650 s,the message fl ow of situation awareness sharing and collaborative decision increases quickly,while the increasing of assurance intelligence message slows down,which means the core of the system structure shifts from the collaborative observation loop to the collaborative decision loop gradually. When the simulation time exceeds 1 250 s,the message fl ow of operation instruction,order and scheme increases rapidly,meanwhile,message of assurance intelligence and situation awareness sharing reaches a steady stage,which indicates that the core of the system structure shifts from the collaborative loop to the collaborative command and controlloop.

Fig.12 Information flow in different loops over simulation time

This experiment fi nds out the core of the system structure and its shifting phenomenon by calculating CNE of different loops and studying fl ow changing of message types.On one hand,this experimentveri fi es the theory that the core of the system structure shifts rapidly as the operation proceeds;on the other hand,it indicates thatthe key point of system defense should make proper shifting and emphasis.Furthermore,in the environmentof cyber combat,it could provide basis for fi nding out the core of the warfare system ofthe enemy and forgiving deadly blow to adversary’s system.

5.2 Efficiency advantages of the adaptive evolvementTo analyze the advantages brought by structure adaptive evolvement to system ef fi ciency,two experiments under the same parameter con fi guration and scenario are carried out.In the fi rstexperiment,there is no mechanism of adaptive evolvement,and the system structure will not be adjusted as the mission,environment and damaging status of units change;while in the second experiment,adaptive evolvement is applied.During the two simulation experiments,the indexes of system ef fi ciency including the quality of targets detecting,targets assigning and targets intercepting in the process ofcombatare estimated respectively and the experimentresults are shown in Fig.13.

As shown in Fig.13(a),when the simulation time is less than 500 s,the mechanism of collaborative observation loop evolution could effectively improve the integrity of target detection;thereby increase the warning time of thesystem towards threat targets.In Fig.13(b),if expected value of target coverage coef fi cient is 3,the observation loop evolution could improve the rate of target coverage. Considering Fig.13(a)and Fig.13(b)together as a whole, when simulation time exceeds 500 s,the system without a dynamic evolution mechanism of C4ISR structure could also detect all the targets,but the target coverage coef fi cient could hardly satisfy the requirement of combat,and the detecting ability of the system is relatively weak.Therefore,some observation units being disturbed or invalid would have a strong impact on the quality of target detection.As shown in Fig.13(c)and Fig.13(d), operational effectiveness is signi fi cantly improved by the C4ISR structure evolution,with the targetassignmentrate and targetinterception rate increasing about10%and 15% respectively.In the scenario of experiment,the locations of some adversary targets are beyond the power of the initial C4ISR system for intercepting;while the mechanism of structure evolution could bring in new decision units and action units to meet the mission requirement to improve the operationaleffectiveness.

Fig.13 Contrastofsystem efficiency in the situations with and without the evolvement mechanism

The experiment concludes that adaptive evolvement of the C4ISR structure could assure target detection quality, assignmentrate and interception rate,especially when the operation is beyond the ability of the initial system,since new proper units could join to enhance the ability of the initialsystem via the mechanism of adaptive evolvement.

5.3 Characteristics of complex network

During the experiment,we count degree distribution and average path length of the system structure.The cumulative degree distribution function P(k)is adopted for analyzing the degree of unit,and P(k)is de fi ned as the proportion of the number of units having the degree equalto or larger than k in the system.Topologicalstructures and complex network coef fi cients over time during the adaptive dynamic evolution of the system structure are shown in Table 1.

As shown in Table 1 and Fig.14,when the system carries out an operation,new intelligence acquisition,decision and action units will join the system structure to satisfy the changing need of target detection,target assignment and interception.As the combat proceeds to destroy invading targets successively,some action units exit the system structure gradually.During the experiment,the degree distributions of the system structure over time all obey the power-law distribution approximately.On the other hand,compared to the number of system units,the value of average path length is relatively small.The experiment results indicate that information age C4ISR structure basically satis fi es the statistical property of scale-free network.In the scale-free network,the numberofunits with high degree is small,while the number of units with low degree is large. Ithas relatively strong robustness forrandom attacks while shows certain vulnerability to deliberate attacks.Accordingly,destroy-resistantmethods of redundant backup and key protection on some key units with high degree in the system could be adopted to improve the robustness of the system structure.

Table 1 Relative parameter of system structure in the processing of dynamic evolvement

Fig.14 Cumulative degree distribution ofsystem structure at different time

6.Conclusion

This paper proposes the information age C4ISR structure modelincluding four kinds of system units and four kinds of relationships among them.It de fi nes the concepts of the collaborative observation loop,collaborative decision loop,command and control loop and collaborative action loop,analyzes the adaptive evolvementprocess and mechanism ofthe system structure in differenttask phases(from sensing a group of targets to attacking those targets),and establishes the adaptive evolvementmodel.With a case of area jointair defense C4ISR,this paper illustrates the ef ficiency superiority of adaptive evolvementand core shifts, and veri fi es the small-world and scale-free attributes ofthe structure in the process of adaptive evolvement.Studying the characteristics of the system structure adaptive evolvement gives an insight into the ability of information age C4ISR to address differenttasks and environmentand provides scienti fi c basis for reasonable system construction, operation managementand system protection.

[1]A.K.Cebrowski,J.J.Garstka.Network-centric warfare:it’s origin and future.US Naval Institute Proceedings,1998.

[2]D.S.Alberts,J.J.Garstka,R.E.Hayes.Understanding information age warfare.USA:CCRP,2004.

[3]D.S.Alberts.The agility advantage:a survival guide for complex enterprises and endeavors.USA:CCRP,2011.

[4]D.S.Alberts.Agility,focus,and convergence:the future of command and control.C2 Journal,2007,1(1):1-30.

[5]Y.S.Lan.Understanding the network-centered command information system.Command Information System and Technology,2010,1(1):1-4.(in Chinese)

[6]Y.S.Lan,F.Ding,H.Wang,etal.Military information infrastructure in information age.Beijing:Military Science Press, 2012.(in Chinese)

[7]S.J.Mao,K.B.Deng,H.Wang,et al.Complexity of netcentric and service-oriented C4ISR system.Command Information System and Technology,2012,3(4):1-6.

[8]G.L.Yang,B.X.Xiu,W.M.Zhang,etal.Dynamic SoS analysis using structure reorganization methodology.Proc.of the 12th IEEE InternationalConference on Information Reuse and Integration,2011:324-335.

[9]M.Jobidon,A.Labrecque,I.Turcotte,et al.Adaptability in crisis management:the role of organizationalstructure.Proc. ofthe 18th International Command and ControlResearch and Technology Symposium,2013:1-19.

[10]Z.Liu,J.C.Huang,Y.J.Tan,et al.Performance-oriented adaptive design for complex military organizations.Proc.of the International Conference on Performance,Safety and Robustness in Complex System and Application,2011:1232-1245.

[11]A.W.Potts,J.W.Kelton.The need for dynamic airspace management in coalition operations.C2 Journal,2011,5(3): 607-621.

[12]A.H.Dekker.Applying social network analysis concepts to military C4ISR architectures.Connections,2002,24(3): 93-1-3.

[13]A.H.Dekker.C4ISR architectures,social network analysis and the FINC methodology:an experiment in military organizational structure.DSTO Electronics and Surveillance Research Laboratory,2002.

[14]Y.S.Lan,K.Yi,H.Wang,et al.Delay assessment method for networked C4ISR system architecture.Systems Engineering and Electronics,2013,35(9):1908-1914.(in Chinese)

[15]G.L.Yang,J.C.Huang,W.M.Zhang,et al.Combat system quantitative assessment method based on FINC model.Computing Engineering,2011,37(10):287-290.(in Chinese)

[16]J.C.Huang,W.M.Zhang,G.L.Yang,et al.The modeling and ef fi ciency analysis method of C2 system ofsystems based on FINC model.Proc.ofthe InternationalConference on Machine Learning and Cybernetics,2010:2026-2030.

[17]S.C.Chang,L.X.Li.Continuous core/periphery architecture features of C2 organizations analysis based on socialnetwork analysis.Proc.of the IEEE International Conference on Computer Science and Automation Engineering,2011:1123-1135.

[18]D.Scheidt,K.Schultz.On optimizing command and controlstructures.Proc.of the 16th International Command and ControlResearch and Technology Symposium(the bestpaper), 2011:1-12.

[19]K.Schultz,D.Scheidt.An application of information theory to command and control.Proc.of the 17th InternationalCommand and Control Research and Technology Symposium, 2012:1-24.

[20]E.Jensen.Operationalizing C2 agility.Proc.of the 17th International Command and Control Research and Technology Symposium,2012:1-16.

[21]R.E.Bryant.C2 in underdeveloped degraded and denied operational environments.Proc.of the 18th International Command and Control Research and Technology Symposium, 2013:1-12.

[22]J.McFarland,D.McConnell,H.Reed,et al.Modeling C2 agility to meetthe demands of a distributed force.Proc.ofthe 18th International Command and ControlResearch and Technology Symposium,2013:1-17.

[23]R.Oosthuizen,L.Pretorius.Assessing command and control system vulnerabilities in underdeveloped,degraded and denied operationalenvironments.Proc.of the 18th International Command and Control Research and Technology Symposium, 2013:1-17.

[24]O.Cetinkaya,Y.Yildirim.Architecture concerns of time sensitive targeting(TST)and NATO TST tool.Proc.of the 18th InternationalCommand and ControlResearch and Technology Symposium,2013:1-24.

[25]D.Essens,M.Spaans,W.Treurniet.Agile networking in command and control.C2 Journal,2007,1(1):211-232.

[26]F.Osinga.Science,Strategy and War:The Strategic Theory of John Boyd.Boston:Eburon Academic Publisher,2005.

[27]B.Brehmer.The dynamic OODA loop:amalgamating Boyd’s OODA loop and the cybernetic approach to command and control.Proc.ofthe 10th InternationalCommand and ControlResearch and Technology Symposium,2005:1-10.

[28]K.Wheeler.Defense information infrastructure(DII)common operating environment(COE)integration and runtime speci fi cation.version 4.1.http://dod.ead.mont.disc.mil/ docs/delivery,1998.

[29]Department of Defense.Global information grid architectural vision 1.0.http://cionii.defense.gov/docs/GIGArch/ GIGArchVision.pdf.

[30]J.R.Cares.Distributed networked operations-the foundations of network centric warfare.New Port:Alidade Press,2006.

[31]C.E.Shannon.A mathematicaltheory ofcommunication.Bell System Technical Journal,1948,26(3):379-423.

Biographies

Yushi Lan was born in 1954.He received his B.E.degree in communication and television from Shandong University in 1982.He is currently a professorate senior engineer of the 28th Research Institute of China Electronics Technology Group Corporation and a Ph.D.supervisor in Nanjing University of Aeronautics and Astronautics.He is also appointed as the director of Chinese Institute of Electronics,vice-director commissioner of the branch Institute of Electronic System Engineering and the chief of editing committee of the Journal ofCommand and Information System.His research interests mainly include military information system integration,military information infrastructure and cyberspace.

E-mail:yushi lan@126.com

Kebo Deng was born in 1980.He received his Ph.D. degree in communication and information system from Nanjing University of Science and Technology in 2009.He is currently a seniorengineer ofthe National Key Laboratory of Science and Technology on C4ISR.His research interests focus on the modeling,simulation and assessmentofthe military information system.

E-mail:dkb612@163.com

Shaojie Mao was born in 1963.He received his B.E.degree in computing mathematics from Nanjing University in 1982.He is a professorate senior engineer of the National Key Laboratory of Science and Technology on C4ISR,and a standing director of Chinese Association for System Simulation.His research interests cover information system integration,system simulation and complex network.

E-mail:maoshaojie@gmail.com

Heng Wang was born in 1977.He received his Ph.D.degree in computer application from Nanjing University of Science and Technology in 2005.He is currently a professorate senior engineer of the National Key Laboratory of Science and Technology on C4ISR.His research interests include information grid,information system integration and complex network.

E-mail:wanheng@gmail.com

Kan Yi was born in 1981.He received his Ph.D.degree in information network from Nanjing University of Posts and Telecommunications in 2010.He is currently a senior engineer of the National Key Laboratory of Science and Technology on C4ISR. His research interests focus on military information system architecture and integration.

E-mail:yikancn@gmail.com

Ming Lei was born in 1983.He received his M.S.degree in probability theory and statistic from Wuhan University in 2008.He is currently an engineer of the National Key Laboratory of Science and Technology on C4ISR.His research interests focus on the modeling and simulation of the military information system.

E-mail:001leiming@163.com

10.1109/JSEE.2015.00036

Manuscriptreceived on March 12,2014.

*Corresponding author.

This work was supported by the National Defense Basic Research Program of China and National Defense Pre-Research Foundation of China.