SUBSTANTIATION OF EXPEDIENCY OF THE COMPLEX APPROACH FOR SUPPLY CHAINS MANAGEMENT IN THE COVID-19 CONDITIONS ROLE OF THE AIRLINE AS A LOGISTICS PROVIDER IN THE

necessary to develop special methodological tools, and, on their basis, software to facilitate decision-making on the timely formation of liquid consignments at the terminal sections of the chain of liquid goods. The article describes a fuzzy network of CLSC in the form of a hierarchical two-level nested Petri net (NPN), the upper level of which reproduces the process of functioning of the focal company as the central element of the CLSC, and each of the components of the lower level of the network model is an elementary Petri net reflecting the logistic processes at the terminal sections of the CLSC production of raw materials and sale of finished products. This article also gives a description of the procedure for creating a fuzzy network model for representing information about business processes that take place during the functioning of the CLSC, taking into account the existing time and resource constraints, in the form of a NPN, which is expanded by introducing fuzzy and temporal statements. Special methods of automated decision- making on the sustainable functioning of CLSCs are described and justified in terms of making a choice on the transport mode and optimal routing. Based on the developed methodological tools, the process of forming and providing the stable functioning of the CLSC for typical food products of the grocery group, namely dried fruits, is considered.

These circumstances determine the insufficient effectiveness of the existing means of information support for the CLCS, and necessitate their modernization, by expanding the concept of SCM (supply chain management), by supplementing knowledge with oriented methods, to achieve a PLM Software (Germany), Unigraphics (USA), etc. Along with this, the specifics of the CLCS functioning does not allow the direct using of the corresponding standard means of automating the processes of interaction among the participants in the chain, since these developments do not provide an effective solution to the entire complex of tasks of information support of processes within the entire chain. In addition, the presence of a theoretical basis in the form of elements of the theory of Petri nets, the theory of temporal statements and fuzzy mathematics makes it possible, through theoretical generalization, to create a methodological basis for organizing transportation between the elements of the CLCS in the "just-in-time" mode. Based on this, there is a need for further development of methods and means of information support for CLCS management processes, in terms of ensuring transportation, in order to create a special applied information technology. Thus, increasing the efficiency of PCP functioning by ensuring the timeliness of deliveries within the chain, through the development and implementation of information support technology for the transportation of goods in terms of ensuring the "just-in-time" mode, is an urgent scientific task to be solve for achieving a par5ticular benefit in application. The purpose of the article is to outline an approach to the rational organization of business processes associated with the most important aspect in the life cycle management of the CLCS -namely ensuring efficient transportation of goods within the entire supply chain.
Problem Statement. Improving the efficiency of cargo transportation within the framework of the CLCS based on the methods and means of artificial intelligence is a complicated and complex problem. One of the options for solving this problem is to implement a set of the following particular tasks consistently.
1. Develop a network model of the complete supply chain that would adequately reflect the hierarchy of the chain, namely its upper level (a focal company for the processing of raw materials) and lower levels reflecting the activities of suppliers of raw materials and finished products distributors.
2. Develop a method for representing time dependencies between business processes in the complete supply chain, which would provide an opportunity to identify deviations in the functioning of the chain and assess the extreme values of these deviations in order to comply with the principle of "just-in-time".
3. Synthesize a method for making decisions on the choice of a optimal transportation route within a complete logistical chain, which would provide an opportunity to reduce the level of uncertainty on time and financial costs in the supply chain operation.
The solution of the above tasks will make it possible to develop an intelligent technology of information support for the functioning of the complete logistics supply chain in terms of organizing cargo transportation.
Basic material and results. Effective supply chain management is not possible without analyzing them at various levelsstrategic, tactical and operational [1].
At the strategic level, the tasks of designing the CLCS and determining the size of service facilities are solved, taking into account international, national and regional features of the development of transport systems. As part of the CLCS, the main terminals, distribution centers, consolidation warehouses are determined, between which regular transportation of raw materials and finished products is carried out (main routes). Other objects in the CLCS are served using a variety of secondary transportation routes.
Based on demand forecasting, the tasks of purchasing and distributing within the service network are solved, taking into account the urgency of supplies, the range of supplied raw materials and the distribution of finished products, the seasonality of production and sale of food products, and the level of transport costs in the supply chains. Based on the amount of transport costs, the problem of determining tariffs for transport services is solved, taking into account the "price / quality" ratio and the dynamics of the use of rolling stock.
At the tactical level of providing transport and logistical services, the plans for the transportation of goods are adjusted taking into account the Bullwhip effect, uneven demand, the presence of rolling stock at the nodes of supply chains. At this stage, based on the chosen strategy for distributing products through sales channels, the scheduling of product delivery by main and subordinate routes is carried out, taking into account the frequency of service, the capacity of warehouses and terminals, and the compatibility of the transported products.
Network management methods can be used while developing models of supply chains, [2][3][4]. Figure 1 shows the structure of a typical CLCS. The focal company (node # 8) receives material resources from three suppliers (nodes # 1, 3, 7). The first supplier works through intermediaries, the third supplier works through intermediaries and directly, The electronic scientifically and practical journal 26-39 "INTELLECTUALIZATION OF LOGISTICS AND SUPPLY CHAIN MANAGEMENT", v.5 (2021) ISSN 2708-3195 https://smart-scm.org the seventh supplier carries out direct deliveries. The focal company (node # 8) uses both direct and indirect channels to market finished products. Nodes ## 9, 10, 11 and 14 represent sales resellers, and Nodes ## 12, 13 and 15 represent retailers. When analyzing the functioning of the CLCS using network models, the concepts of "central node" of the network and "subnet" form the foundation. The central node of the network is the focal company. A subnet is a part of a network, a collection of connected links. The left and right subnets in the network are selected. The left subnet is formed by the central link and all the links that are involved in the supply of raw materials (primarily fresh fruit) to the focal company. The right subnet is formed by the central link, sales intermediaries and retailers. The development of an effective CLCS involves solving a set of tasks related to minimizing logistical costs for both the focal company and other participants in the chain. To formalize the problem under consideration, we will separately consider the planned indicators of purchases, sales, and costs for the left and for the right subnets of the CLCS. Let the size of the production program of the focus company be equal to the plan for selling products by retailers. Assume that the prices on the purchased raw materials and sold finished products are stable [5].
The planned demand of the focal company for the raw materials is: , where R li -consumption rate of l-th type of raw material for production of the i-th type of the product; mproduct range of the focal company moved within the logistical chain; Qiproduct volume required to fulfil the production plan of the focal company.
The planned volume of products manufactured by the focus company is determined based on the production (sales) plan, taking into account a set of factors that , where Pjproduction plan (sales) of the j-th product ; kijvolume of the i-th type of raw material used in production of the j-th type of product; nnumber of items in the range of sold products. Logistic costs associated with the procurement of raw materials (costs of the left subnet of the supply chain): , where Zlpdelivery costs of the l-th type of raw materials from the p-th supplier; unumber of suppliers; trange of supplied raw materials.
The sales plan is determined based on research of target market segments: , where Pjksales plan of the j-th type of product for the k-th segment; sthe number of segments.
Delivery costs associated with the sales of finished products (costs in the right subnet): , where Z jkdelivery costs on the j-th type of product in the k-th segment.
The objective function, which assumes minimization of the total logistical costs of the focal company related to procurement and sales, is: → .
The solution to this problem is the selection of suppliers of raw materials and the volumes of these supplies, as well as the selection of links in the distribution network and the distribution of batches of finished products among them.
The process of creating, deploying and supporting the functioning of a CLCS can be adequately represented in the form of a hierarchical two-level nested Petri net framework (NPN) [6]: https://smart-scm.org reflects a specific production situation because the focal company has the finished products in the warehouse. The NPN class gives an opportunity to reflect the hierarchical structure of the CLC in an adequate manner. At the same time, there is no mechanism for reflecting time in this kind of Petri nets. Based on this, in order to comply with the 'just-in-time' principle, it is critically important to expand the model presented in the form of NPN in the aspect of taking into account time dependences.
To build a model for the explicit representation of time during the functioning of the CSC, it is necessary to implement the following steps [7]: 1. Select base primitives of time and define base relationships between them 2. Introduce the necessary elementary functions for transforming primitives and relations.
3. Represent the properties of the structure of time using axioms that determine the basic properties of time and the properties of basic relations.
4. Describe a way to represent time dependencies.
5. Choose a method to associate logical statements with time.
6. Synthesize the theory of temporary statements.
At the first stage of the knowledge model synthesis, taking into account time dependencies, moments and / or time intervals are used as basic primitives. If necessary, time constants are used to indicate the moments of time and intervals (seconds, minutes, hours, days, dates, time, etc.), besides, the primitive "Duration" is used, with the help of which the distance between the moments of time is set up.
The second stage: some of the functions are built on the basic relations and are their functional version; the functions allow converting between temporary primitives.
At the third stage of the model synthesis, it is necessary to specify the connection among several primitives of time in the form of the corresponding axioms, taking into account the properties (discreteness / continuity) of time, characteristics of individual elements of the CLSC; at the fourth stage, a method for describing time dependences (time model) is chosen.
The fifth stage of the synthesis of the model of time dependencies involves the use of approaches that have good expressiveness.
At the last, sixth stage, it is necessary to define the basic temporary statements, and set up their properties using a set of axioms.
The information about the current state of tasks scheduled for execution is updated in the model  The timeliness of the implementation of tasks is characterized by the attribute Timeliness, which determines task affiliation of In order to activate the The purpose of the developed method is to identify deviations from the schedule when performing tasks during the functioning of the CLCS and to formulate instructions for stabilizing the current state of the chain.
The first stage of the method (steps 1-4) is carried out in parallel, asynchronously and cyclically during each cycle of monitoring the links of the logistics chain, this makes it possible to take into account the current state of its individual links in the process of supporting decision-making on the implementation of CLCS.
The second stage of the method (steps 5-7), in the case of time delays, makes it possible to form solutions to stabilize the CLCS functioning.
The initial state (CLCS start) is assigned as follows: 8. The execution of the method will end under the conditions if the current state of all tasks has the status "Completed", that is, the functioning of the CSC is completed.
To determine the mode of transport for the dried fruits delivery, it is necessary to take into account the six main factors influencing decision-making process: delivery time; transportation costs; compliance with the delivery schedule; frequency of departures; variety of good to be shipped, and the ability to deliver cargo anywhere. In the process of purchasing and delivering material resources within the CSC, as well as distributing finished products to consumers, a focal company can use various types of transport, various logistical partners and various transportation options [6].
The selection of the optimal mode of transport for the formation of the CSC will be carried out while considering that the delivery time and costs are the optimization criteria. The choice of the type of transport will be carried out using artificial intelligence methods, namely fuzzy modeling.
Let us make a choice of the mode of transport for the supply of raw materials and supplies from raw material suppliers to the focal company (the left side of the network). For this purpose, we construct a fuzzy model [8] based on two binary fuzzy relations S and T. The first of these fuzzy relations is built on two basic sets X and Y, and the second -on two basic sets Y and Z. Here X describes the set of modes of transport by which the transportation can be carried out, Ythe set of transport options, and Ztransport characteristics. The fuzzy relation S meaningfully describes the relationship between the mode of transport and the transportation option, and T describes the b) y 1 -«unimodal», y 2 -«mixed», y 3 -«combined», y 4 -«intermodal», y 5 -«terminal», y 6 -«multimodal»; b) z 1 -«delivery time», z 2 -«delivery frequency», z 3 -«schedule reliability», z 4 -«handling variety of cargo», z 5 -«ability to deliver goods in any geographical point», z 6 -«delivery costs».
As a result of the analysis of the calculated values of the membership function, the best option according to the criteria "Delivery time" and "Delivery cost" will be the use of road transport, since the membership functions are equal to µS*T(<x3, z1>)=0,8, µS*T(<x3, z6>)=0,6, respectively.
The solution to the problem of choosing a mode of transport does not in itself ensure the efficiency of the processes of dried fruits transporting within the CSC framework; the quality and speed of transportation is directly influenced by the optimal choice of the route. The safety of the cargo and the extraction of the actual maximum profit is achieved by drawing up an optimal route [5]. When drawing up an optimal route, it is necessary to take into account the location of the final point of delivery, the dimensions and weight of the cargo, as well as its characteristics. Taking into account the listed parameters, a necessary transportation vehicle is selected.
While designing the CLCS, a route which takes into account all the driver's possible stoppage places for meals and overnight accommodation as well as the customs control points is drawn up. In addition, it is necessary to take into account the state of the road surface and the time required for crossing the borders of other states as well as the characteristics of each region on the route of the cargo transportation. Also it is necessary to take into account the road's width. The quality of the pavement's surface and the weather conditions that affect the conditions of the road. To ensure just-in-time delivery, the speed limits on certain segments of the route must be considered. In order to choose the optimal transportation route, it is recommended to build a fuzzy model in the MATLAB tool environment and develop an expert system while functionality will be based on fuzzy inferences. The interactive mode is provided by using the Fuzzy Logic Toolbox package, which is a part of the MATLAB environment [9].
Four fuzzy linguistic variables must be used as input parameters of the fuzzy inference system to determine the rational transportation route: weather conditions; the quality of the pavement surface; the number of speed limits encountered; time required for passing the customs post. In this case, the output variables will be: transportation time and transportation cost.
The Mamdani method is used as a fuzzy inference scheme; therefore, the activation method will be MIN [8]. The center of gravity method was used as a defuzzification method for the obtained result.
To build a fuzzy model for choosing an optimal transportation route, it is assumed that all considered input variables are measured in points in the range of real numbers from 0 to 10, where the lowest estimate of the value of each of the variables is 0, and the highest is 10.
The problem of fuzzy modeling was solved based on the Mamdani rule, while the parameters of the developed fuzzy model were proposed by MATLAB on default were the following items remained unchanged: logical operations (min for a fuzzy logical "AND", max for a fuzzy logical "OR"), implication method (min), aggregation method (max) and defuzzification method (centroid). After that, while solving the problem, the membership functions of the terms were determined for each of the four inputs and one output variable of the considered fuzzy inference system. Further, for the developed expert system, a knowledge base of 30 rules was formed. Figure 2 shows the editor of the rules included in the fuzzy knowledge base of the expert system, called by the ruleedit ('marschrut') function. Then the analysis of the constructed fuzzy inference system for the considered problem of choosing a rational route for the transportation of a batch of dried fruits along the international route is carried out. By entering the value of the input variables for the first option of the route, the value of the input variable "weather conditions" is set up as 5 points, the value of the input variable "Pavement quality" is 4 points, the value of the input variable "speed limit" is 6.8 points, and the value of the input variable "The customs post passing" is 7 points.
As a result, the fuzzy inference procedure implemented using the MATLAB system for the developed fuzzy model, gives the value of the output variables "Delivery time" and "Transportation cost", equal to 41.5 hours and 12.7 thousand UAH respectively.  For the second route, the value of the input variable "Weather conditions" was estimated as 5 points, the value of the input variable "Pavement quality" as 7 points, the value of the input variable "Speed limit" as 4.5 points, and the value of the input variable "The customs post passing" as 3.5 points. As a result, the fuzzy inference procedure made it possible to obtain the values of the output variables "Delivery time" and "Transportation cost" equal to 49.2 hours and 15.6 thousand UAH respectively. Obviously, based on the results obtained for this example, it is more profitable to transport goods within the framework of the CSC using the first route option.
The considered fuzzy model produces sufficiently adequate results. However, for achieving a higher level of precision, it is necessary to use additional rating methods of individual quantitative values for input and output linguistic variables. In practice, to solve this problem, it is reasonable to use such tools that would be highly transparent and visual and would facilitate an interface with the software system without having advanced knowledge and skills in the field of software engineering. Figure 3 shows the visualization process of the fuzzy inference surface of the considered model for the input variables "Speed limits" and "The customs post passing". This visualization tool facilitates establishing the relationship between the output variable values and the individual input variables values of the fuzzy model. The analysis of these relationships serves as the basis for changing the membership functions of input variables or fuzzy rules in order to improve the adequacy of the fuzzy inference system.