Classification of DTIPs

     In subsection "What is a discrete process" of this section of the website is given the classification of discrete processes on three signs: controllability (controlled and non-controlled), predictability of behavior (random and deterministic) and duration (infinite, limited in time and instant). Based on this general classification, divide the class of discrete technological and information processes (DTIPs) into different categories, analyzing in such case their influence on the mechanism of applying IT AC DTIP.

     First, we note that all DTIPs belong to the category of controlled discrete processes. This means that for each such process it is possible to change a control strategy, applied to its objects, in order to improve the effectivenessof its functioning. Actually, DTIP is represented as a controlled service system, since this system provides for changing of a control strategy.

     This raises the following question: by what elements of DTIP (parameters of its objects, procedures, etc.) we can control it? After all, the more of these controlled elements, the more effective may be a control strategy for this process, synthesized using Technology.

   Consider an example confirming the above. Let DTIP is a process of processing on a machine with changeover of a stationary flow of different-type details. A quality criterion here is the average value per time unit of losses from delays in processing details, as well as from denials of their processing. Consider two options for specifying controlled elements of DTIP. In the first case it is allowed to change the order of entry of details to the machine after its release from the previous ones, and in the second — to change this order in any moment of time (including during processing), and throw away any unprocessed details. It is easily seenthat that in the latter case it will be possible to control this DTIP more effectively, especially when the flow of incoming details will be too large.

     In each particular case a set of DTIP elements, which are available for control, is specific (individual). Therefore,there is nosense todraw up a list of categories of DTIPs on the signs of control, becausesuch categoriescan be quite a lot. In this connection, we list only the most characteristic elements of DTIP, by which in certain cases we can control. These include the following:

  • order of entry of demands (service stages of network demands) to devices of service;
  • order of leaving the system by not fully serviced demands;
  • modes (speeds) of serving demands by devices.

     We now turn to the classification of DTIPs on the sign of predictability of behavior. It will be the same as for any discrete processes (see above). In other words, on this sign the DTIPs are divided into two categories: random (probabilistic) and deterministic. In order to understand, to which of these categories is related a specific DTIP, you need to know those parameters and characteristics of its objects (demands, devices and mediums), which influence its functioning. If all of them are deterministic, hence the process is the same. Otherwise, it will be random.

     Depending on whether DTIP is a random or deterministic process, can be applied different procedures of Technology to solve a task of its numerical optimization. Besides that, the output result, which is a synthesized tabular strategy of control (TSC), will also be used differently. The control of a random process takes place with the help of a situational manager, that automatically analyzes a current state of the process and retrieves from TSC a corresponding control.

     For deterministic process the control procedure can be realized not only by the above-mentioned automatic way (in the mode of situational dispatch), but also by generating a list of controls or graphiccharts. This is due also to the fact that at the same number of phase states for both processes: deterministic and random, a number of actually used states, in which occur transitions, for the first process will be much lessthan for the second one.

     Considering the above, we can make the following two conclusions:

  • before applying Technology to some DTIP, you must find out whether it is random or deterministic;
  • if an initial DTIP have some random parameters, and when artificial replacing them by deterministic ones the functioning of the process does not change (concerning the value of its quality criterion), then it is desirable to make such a change for the above reasons.

Note. The abovementioned replacement can be performed in case, when a random time of serving a demand with its average value T will be replaced by a fixed service time, equal to T, the system has one device of service and interruptions of serving demands are absent. It is easy to show that under such a replacement the functioning of DTIP remains unchangedfor any control strategy.

Classes of DTIPs of different duration

     Now classify DTIPs on the sign of their duration. In contrast to the general classification of discrete processes on this sign (infinite, limited in time and instant), for DTIPs is one clarification: the infinite processes can be only stationary. In other words, depending on the flow time of DTIPs they are divided into three classes of processes: stationary (class DTIP-S), time-limited (class DTIP-T) and instant (class DTIP-I).

     Let's characterize processes of these classes, bringing them demonstrative examples.

     The stationary is such DTIP, which has the following three properties:

  • its all parameters and characteristics, as well as applicable controls don't depend on the time of observing the process;
  • DTIP has at least several return states to which happens regular transitions during the process flow;
  • actual duration of DTIP should be large enough, that during this time there have been repeated transitions to each of the overwhelming number of its recurrent states.

Here are some examples of stationary DTIPs:
1)   process of transferring discrete messages in a communications network;
2)   process of repairing faulty objects of urban infrastructure;
3)   process of functioning an emergency city service.

    To time-limited DTIPs are related the processes of network planning and scheduling, which are widely encountered in real life. Each such process has the following two properties:

  • it lasts limited and non-zero lapse of time;
  • it is characterized by a constant number of objects (incoming and outgoing flows of demands and devices are absent here).

     The following are typical examples of time-limited DTIPs:
1)   process of designing or construction of some building;
2)   process of creating a new product;
3)   processing a batch of details by a groupof machines.

     Under an instant DTIP is understood a so-called process of selecting choices, for which the time factor is absent, in connection with which it can be considered instant. In this process a device of service, that is purposed for the group service of demands, is a some resource, specified by you (allocated amount of money, a volume of some capacity, allowable load weight, etc.), and demands — samples of either acquired goods (services), or transported cargo, selected from their specified set. 

     Typical examples of instant DTIPs are:
1)   process of investing funds into sources of income;
2)   process of spending a given amount of money for purchase of goods or services;
3)   process of drawing up budget of a Ministry or department