Resource allocation method and full duplex communication system

Abstract

The present disclosure provides a resource allocation method. The resource allocation method includes the following steps: selecting multiple first selected virtual nodes according to multiple virtual pheromonal trails on multiple virtual edges, in which the first selected virtual nodes forms at least one virtual tour, and the virtual tour includes multiple first virtual edges; updating the virtual pheromonal trails on the virtual edges according to virtual distances corresponding to the first virtual edges of the virtual tour; selecting multiple second selected virtual nodes according to the updated virtual pheromonal trails on the virtual edges, in which the second selected virtual nodes form at least one resulting virtual tour; allocating the resource blocks to selected user pairs according to the resulting virtual tour.

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present disclosure relates to a resource allocation method and a full duplex communication system, and more particularly, to a resource allocation method and a full duplex communication system capable of allocating wireless resources in a multi-cell scenario.

2. Description of the Prior Art

As the demand of wireless service increases, the utilized frequency spectrum is getting crowded, which might degrade the quality of service (QoS) of wireless systems. Enhancing data rate is always a goal for the next-generation mobile communication. Full-duplex (FD) communications, allowing simultaneous transmission and reception on the same frequency carrier(s), attracts more attentions recently, which is expected to be a promising way to increase spectrum efficiency.

Previously, strong self-interference makes FD communications difficult to be realized. Thanks to the breakthroughs in hardware development, self-interference is able to be reduced by 110 dB, which makes FD communications possible to be realized and being able to upgrade the capacity to a new level. Therefore, resource allocation for the FD system is critical.

SUMMARY OF THE INVENTION

It is therefore a primary objective of the present disclosure to provide a resource allocation method and a full duplex communication system capable of allocating wireless resources in a multi-cell scenario.

The present disclosure provides a resource allocation method applied in a full duplex communication system. The full duplex communication system includes multiple stations and multiple user devices, and the full duplex communication system operates on multiple resource blocks. The resource allocation method includes the following steps: selecting first selected virtual nodes according to virtual pheromonal trails on virtual edges, in which the first selected virtual nodes forms at least one virtual tour, and the virtual tour includes first virtual edges; updating the virtual pheromonal trails on the virtual edges according to virtual distances corresponding to the first virtual edges of the virtual tour; selecting second selected virtual nodes according to the updated virtual pheromonal trails on the virtual edges, in which the second selected virtual nodes form at least one resulting virtual tour; and allocating the resource blocks to selected user pairs according to the resulting virtual tour, in which selected user pairs correspond to the resulting virtual tour. The user devices and stations form user pairs, and each user pair includes a downlink user device and an uplink user device, among the user pairs, along with a station among the stations. The user pairs represent virtual nodes, and the first selected virtual nodes are among the virtual nodes. A virtual edge, among the virtual edges, is formed between two virtual nodes among the virtual nodes. The virtual distances are related to data rates.

The present disclosure further provides a full duplex (FD) communication system. The full duplex communication system includes stations, user devices, and a computing device. The full duplex communication system operates on multiple resource blocks. The computing device includes a processing unit and a storage unit. The storage unit stores a program code to be executed by the processing unit to implement the following steps: selecting first selected virtual nodes according to virtual pheromonal trails on virtual edges, in which the first selected virtual nodes form at least one virtual tour, and the virtual tour includes first virtual edges; updating the virtual pheromonal trails on the virtual edges according to virtual distances corresponding to the first virtual edges of the virtual tour; selecting second selected virtual nodes according to the updated virtual pheromonal trails on the virtual edges, in which the second selected virtual nodes form at least one resulting virtual tour; allocating the resource blocks to selected user pairs according to the resulting virtual tour, in which the selected user pairs correspond to the resulting virtual tour. The user devices and stations form user pairs, and each user pair includes a downlink user device and an uplink user device, among the plurality of user pairs, along with a station among the stations. The user pairs represent virtual nodes, and the first selected virtual nodes are among the virtual nodes. A virtual edge, among the virtual edges, is formed between two virtual nodes among the virtual nodes. The virtual distances are related to data rates.

These and other objectives of the present invention will no doubt become obvious to those of ordinary skill in the art after reading the following detailed description of the preferred embodiment that is illustrated in the various figures and drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of a full duplex (FD) communication system according to an embodiment of the present disclosure.

FIG. 2 illustrates schematic diagrams of ants finding food from home.

FIG. 3 illustrates a graphic representation of an FD communication system according to an embodiment of the present disclosure.

FIG. 4 illustrates a virtual path corresponding to a resource block according to an embodiment of the present disclosure.

FIG. 5 illustrates a plurality of virtual paths corresponding to a plurality of resource blocks according to an embodiment of the present disclosure.

FIG. 6 is a schematic diagram of a resource allocation method according to an embodiment of the present disclosure.

DETAILED DESCRIPTION

FIG. 1 is a schematic diagram of a full duplex (FD) communication system 10 according to an embodiment of the present disclosure. The FD communication system 10 comprises a plurality of stations BS, a plurality of user devices UE and a computing device 12. The user device UE may be a phone, a tablet computer, a laptop and the like. The station BS may be a macro base station in a cellular network scenario. The station BS may also be a femto base station or an access point (AP) in a WLAN (wireless local area network) scenario. The plurality of stations BS correspond to a plurality of cells respectively. The cell represents a service area covered by the corresponding station BS. The cells in FIG. 1 are illustrated as hexagons for exemplary purpose, but not limited thereto. The computing device 12 may be a computer or a server, comprising a processing unit 120 and a storage unit 122. The processing unit 120 may be one or more processor(s), e.g., CPU (central processing unit). The storage unit 122 may be a non-volatile memory (NVM), e.g., an electrically erasable programmable read only memory (EEPROM) or a flash memory. The computing device 12 is coupled to the plurality of stations BS and configured to execute a resource allocation process to obtain a resource allocation scheme. Based on the resource allocation scheme obtained by the computing device 12, the plurality of stations BS allocates wireless resources to the user devices UE, to enhance an overall data rate. The storage unit 122 is configured to store a program code 124 to be executed such that the processing unit 120 to execute the resource allocation method.

The FD communication system 10 operates on a plurality resource blocks RB (not shown in FIG. 1). The resource block RB represents a certain frequency spectrum within a certain period of time. Practically, in an OFDM (orthogonal frequency division multiplexing) system, the resource block RB may represent couples of frequency subcarriers or one subcarrier.

In the FD communication system 10, a plurality of user pairs UP is formed. For example, a user pair UP_a represents a downlink user device UE_DL,a and a uplink user device UE_UL,a along with a station BS_b within a cell b. When a resource block RB_c is assigned to the user pair UP_a in the cell b, it means that the uplink user device UE_DL,a in the cell b transmits data packets to the station BS_b (via an uplink direction) using the resource block RB_c and the station BS_b transmits data packets to the downlink user device UE_DL,a (via a downlink direction) in the cell b using the resource block RB_c at the same time. In other words, the plurality of user devices UE and the plurality of stations BS form the plurality of user pairs UP, in which the user pairs UP correspond to the stations BS respectively.

The resource allocation problem for the FD communication system basically refers to determination of which resource block being assigned to which user pair, such that an overall capacity can be maximized. In the present disclosure, the terminologies “capacity”, “data rate”, “transmission rate”, “spectrum efficiency”, and “throughput” are used interchangeably. In an embodiment, the capacity or the spectrum efficiency may be evaluated in terms of bps/Hz (bits per second per Hz), or bps (bits per second).

In the present disclosure, a concept of an Ant Colony Optimization (ACO) algorithm is adopted to solve the resource allocation problem for the FD communication system. A traditional ACO routing technique models the establishment of routes based on the mechanisms, used by ant colonies, of establishing and maintaining paths to desirable food sources. Ants randomly choose a path to a food source. On the chosen path, the ants deposit a chemical substance called “pheromone”, which allows other ants to follow the pheromone scent and hence the chosen path. When more ants traverse a path, the pheromone deposit on that path gets more intense, thereby allowing more ants to become attracted to and use that path to travel to the food source and back. The concentration of the pheromone scent on paths to food sources evaporates after certain amount of time. Thus, longer and less traveled paths lose their pheromone scent faster than shorter and more traveled paths. For this reason, more ants would choose the shorter/shortest paths to food sources.

FIG. 2 illustrates schematic diagrams of ants finding food from home. At the beginning, every ant moves randomly and leaves pheromone on the path through which it passes. Some ants go shorter paths and some go longer, as shown in the subfigure 2a. Next time when the ants leave home, they would sense the pheromone on the path and decide which way to go accordingly, as shown in the subfigure 2b. The higher concentration of pheromone the ants detect, the more probably the ants would follow. On the other hand, the more frequently the path is passed, the more pheromone on that path is retained. Eventually, the ants might find a shortest path, as the subfigure 2c shows.

Specifically, a (, ) graph is used to model the routing of the ants in a conventional ACO algorithm, where represents a set of nodes including N nodes and represents a set of edges including E edges, and the edges are between the nodes. The ACO algorithm operates under three basic assumptions: 1) each ant chooses a node to travel/visit based on a transition probability, which is a function of an edge distance (to the node) and an amount of pheromonal trails left on that edge; 2) each ant would not visit a city twice/again if the tour has not accomplished; 3) the ant lays pheromonal trails on the passed edges after the tour is accomplished. In the present disclosure, the terminologies “pheromonal trail” and “pheromone” are used interchangeably.

Mathematically, during an operation of the conventional ACO algorithm, a transition probability of a k^th ant traveling from an i^th node (or, a node i) to a j^th node (or, a node j) at a time t may be expressed as

$\begin{array}{cc}{p}_{i,j}^k\left(t\right)=\{\begin{array}{c}\frac{{{{\tau }_{i,j}\left(t\right)}^{\alpha }\left[{\eta }_{i,j}\right]}^{\beta }}{\sum _{s\in F_k}{{\left[{\tau }_{i,s}\left(t\right)\right]}^{\alpha }\left[{\eta }_{i,s}\right]}^{\beta }},\mathrm{if}j\in F_k\\ 0,\mathrm{otherwise}\end{array}.& \left(\mathrm{eq}.1\right)\end{array}$

In eq. 1, τ_i,j(t) represents the (virtual) pheromone left on an edge e(i, j) between the node i to the node j at the time t, η_i,j=1/d_i,j represents the visibility (preference) of the edge e(i,j), and d_i,j represents the distance of the edge e(i, j). Parameters α and β are used to reflect the significance of the pheromone and the visibility (preference). The set F_k represents a feasible set including the nodes which are allowed to be visited for the k^th ant, i.e., feasible for the k^th ant, and the feasible set is constrained by the assumption 2 stated above. Once the k^th ant accomplishes its tour, the virtual pheromone τ_i,j(t) on the edge e(i,j) may be updated asτ_i,j(t+n)=ρ·τ_i,j(t)+Δτ_i,j.  (eq. 2)

In eq.2, ρ is an evaporation coefficient, meaning that (1−ρ) of the pheromone τ_i,j(t) would be evaporated between the time t and the time (t+n), n is an amount of time which takes the k^th ant to finish a tour. In other words, ρ·τ_i,j(t) is the remaining pheromone on the edge e(i, j) at the time (t+n), and Δτ_i,j is the newly laid pheromone on the edge e(i, j) between the time t and the time (t+n). The incremental virtual pheromone Δτ_i,j may be expressed as

$\begin{array}{cc}\Delta {\tau }_{i,j}=\sum _{k=1}^K\left(\Delta {\tau }_{i,j}^k\right).& \left(\mathrm{eq}.3\right)\end{array}$

In eq.3, Δτ_i,j^k represents the pheromone laid by the k^th ant on the edge e(i, j) between the time t and the time (t+n), K represents a total number of ants. The incremental pheromone Δτ_i,j^k may be expressed as

$\begin{array}{cc}\Delta {\tau }_{i,j}^k=\{\begin{array}{c}\frac{Q}{L_k},\mathrm{if}\mathrm{the}{k}^{\mathrm{th}}\mathrm{ant}\mathrm{passes}\mathrm{through}\mathrm{the}\mathrm{the}\mathrm{edge}\left(i,j\right)\\ 0,\mathrm{otherwise}\end{array}.& \left(\mathrm{eq}.4\right)\end{array}$

In eq. 4, Q can be a constant representing a total amount of pheromone which is to be laid by the ant during its tour, and L_k is the tour length of the k^th ant.

The conventional ACO algorithm may be presented as Table I. In Table I, N_itr denotes an iteration index, and step, from 1 to n, denotes a step index, meaning that each ant takes n steps to accomplish one tour, where one tour takes n units of time. From Table I, the tours taken by the K ants might converge to a specific tour, meaning that every ant may eventually select the shortest tour to travel, when the iteration index N_itr is less than a maximum iteration limit N_itr,MAX.

TABLE I
1 // Initialization
2 Set t = 0
3 Set Nitr = 0
4 Set τi,j(t) = c for all edges and Δτi,j =0 , for all i,j ∈   , where   is
the set of nodes
5 while (Nitr< Nitr,MAX) or (Lk1 ≠ Lk2, ∀ k1, k2∈ {1,...,K}) do
6  Uniformly Place K ants on N nodes
7   Δt = 0
8  for step = 1 to n do
9    Δt = Δt +1
10    for k = 1 to K do
11     Move the kth ant from node i(step,k) to node j(step,k)
according to eq. 1
12     Update Fk
13    end
14  end
15  for k = 1 to K do
16   Calculate Lk
17   Update pheromonal trails τi,j(t+n) according to eq.2, eq. 3
and eq. 4
18   Update the shortest tour found
19  end
20 t = t + Δt
21 Nitr = Nitr + 1
22 Reset Δτi,j =0 for all i,j ∈
23 end
24 Print Shortest Tour

By properly defining the graph (, ) the ACO algorithm may be applied to solve the resource allocation problem for the FD communication system. The resource allocation problem for the FD communication system is to assign the plurality of resource blocks RB to the plurality of user pairs UP, such that the overall system data rate can be maximized. In the present disclosure, a virtual node represents a user pair UP within a cell. A virtual edge represents a situation/status that two user pairs UP in two cells are assigned/allocated to the resource blocks RB with the probabilities defined in eq. 1.

FIG. 3 illustrates a graphic representation of an FD communication system 30 according to an embodiment of the present disclosure. As shown in FIG. 3, the FD communication system 30 comprises cells CL1-CL3 corresponding to the stations BS₁-BS₃ (not shown in FIG. 3) respectively. The computing device of the FD communication system 30 is also not shown. User pairs (or virtual nodes) UP_{1_1}-UP_{1_2}, UP_{2_1}-UP_{2_3}, and UP_{3_1}-UP_{3_3} are within the cells CL1, CL2, CL3, respectively. The straight lines in FIG. 3 represent edges between the virtual nodes UP_{1_1}-UP_{1_2}, UP_{2_1}-UP_{2_3}, UP_{3_1}-UP_{3_3}. The user pair (or the virtual node) UP_x,y represents a downlink user device UE_DL,y and a uplink user device UE_UL,y, along with the station BS, corresponding to the cell CLx. In an embodiment, FIG. 3 may be interpreted as there are 4, 6, 6 users within the cells CL1, CL2, CL3, respectively, but not limited therein. In another embodiment, FIG. 3 may be interpreted as there are 2, 3, 3 users within the cells CL1, CL2, CL3, respectively. The terminologies “user” and “user device” are used interchangeably in the present disclosure. Note that, the virtual node UP_{x_y} illustrated in FIG. 3 has nothing to do with practical positions of the downlink user device UE_DL,y and the uplink user device UE_UL,y, in the cell CLx.

A virtual edge e(1_1, 2_1) between the virtual nodes UP_{1_1} and UP_{2_1} represents that a specific resource block RB is allowed to be allocated to the user pair UP_{1_1} in the cell CL1 and the user pair UP_{2_1} in the CL2. Since one resource block RB is assigned to only one user pair in one cell, as shown in FIG. 3, there is no edge between virtual nodes within the same cell.

In an embodiment, a virtual edge distance d_{x1_y1,x2_y2} of a virtual edge e(x1_y1, x2_y2) between the virtual node UP_{x1_y1} in the cell CLx1 and the virtual node UP_{x2_y2} in the cell CLx2 may be defined as

$\begin{array}{cc}{d}_{x1\mathrm{\_y}1,x2\mathrm{\_y}2}=\frac{\mathrm{interference}\left({\mathrm{UP}}_{x1,y1}\mathrm{to}{\mathrm{UP}}_{x2,y2}\right)}{\mathrm{Rate}\left({\mathrm{UP}}_{x2,y2}\right)}.& \left(\mathrm{eq}.5\right)\end{array}$

In another embodiment, the virtual edge distance d_{x1_y1,x2_y2} of the edge between the virtual node UP_{x1_y1} and the virtual node UP_{x2_y2} may be defined as

$\begin{array}{cc}{d}_{x1\mathrm{\_y}1,x2\mathrm{\_y}2}=\frac{1}{\mathrm{Rate}\left({\mathrm{UP}}_{x1,y1}+{\mathrm{UP}}_{x2,y2}\right)}.& \left(\mathrm{eq}.6\right)\end{array}$

In eq. 5, the term “Rate (UP_x2,y2)” may be a summation of transmission rates of downlink and uplink corresponding to user pair UP_x2,y2 and the term “interference(UP_x1,y1 to UP_x2,y2)” may be interference caused by the user pair UP_x1,y1 to UP_x2,y2. In eq. 6, the term “Rate (UP_x1,y1+UP_x2,y2)” may be a summation of transmission rates of downlinks and uplinks corresponding to user pairs UP_x1,y1 and UP_x2,y2, respectively. The transmission rate corresponding to each link may be expressed as

$\begin{array}{cc}R={\log }_2\left(1+\mathrm{SINR}\right)={\log }_2\left(1+\frac{\mathrm{signal}\mathrm{power}}{\mathrm{interference}+\mathrm{noise}}\right).& \left(\mathrm{eq}.7\right)\end{array}$

In eq.7, the term “interference” represents an undesired signal power received at the user device UE and/or the station BS, and the term “signal power” represents desired signal power received at either the user device UE or the station BS.

FIG. 4 illustrates how the k^th ant travels between the virtual nodes UP_{1_1}-UP_{1_3}, UP_{2_1}-UP_{2_3}, UP_{3_1}-UP_{3_3}, to allocate a resource block RB₁. At the beginning, as shown in the subfigure 4a, the k^th ant has 9 options, which means that the k^th ant is allocated to one of the virtual nodes UP_{1_1}-UP_{1_3}, UP_{2_1}-UP_{2_3}, UP_{3_1}-UP_{3_3}, as a starting point. The feasible set Fk for the k^th ant is Fk={UP_{1_1}-UP_{1_3}, UP_{2_1}-UP_{2_3}, UP_{3_1}-UP_{3_3}}. The decision of the k^th ant choosing the starting point may be based on a plurality of initial probabilities, which will be described later on.

In an embodiment, the k^th ant may be placed at the virtual node UP_{1_1}, for example, as the starting point (or a first selected virtual node), as shown in the subfigure 4b. It means that the resource block RB₁ may be assigned to the user pair UP_{1_1}. The feasible set for the k^th ant shrinks as Fk={UP_{2_1}-UP_{2_3}, UP_{3_1}-UP_{3_3}} by removing UP_{1_1}-UP_{1_3} in the cell CL1, meaning that the k^th ant has 6 options to move on for the next step.

In an embodiment, the k^th ant may choose the virtual node UP_{2_2}, for example, as the second visited virtual node (or a second selected virtual node), as shown in the subfigure 4c. An edge is formed between the virtual nodes UP_{1_1} and UP_{2_2}, meaning that the resource block RB₁ may have potential to be assigned to the user pairs UP_{1_1} and UP_{2_2}. After the k^th ant chooses the virtual node UP_{2_2}, the feasible set for the k^th ant shrinks as {UP_{3_1}-UP_{3_3}} by removing UP_{2_1}-UP_{2_3} in the cell CL2, meaning that the k^th ant has 3 options to move on for the next step.

In an embodiment, the k^th ant may choose the virtual node UP_{3_3}, for example, as the third visited virtual node (or a third selected virtual node), as shown in the subfigure 4d. The decision made by the k^th ant to choose the next virtual node to visit may be based on a plurality of transition probabilities, which is similar to eq. 1 and will be detailed later on. After the k^th ant chooses the virtual node UP_{3_3}, the k^th ant goes back to the starting point UP_{1_1}, and finishes a first part of its tour, which is corresponding to the resource block RB₁. The virtual path shown in FIG. 4 means that the resource block RB₁ may be assigned to the user pairs UP_{1_1} in the cell CL1, UP_{2_2} in the cell CL2 and UP_{3_3} in the cell CL3.

Specifically, FIG. 5 illustrates virtual paths φ1-φ3 of the k^th ant to allocate resource blocks RB₁-RB₃. The virtual paths φ1-φ3, corresponding to the resource blocks RB₁-RB₃, are illustrated in the subFIGS. 5a-5c, respectively. The subfigure 5d illustrates a combination of the virtual paths φ1-φ3. The virtual paths φ1, φ2 and φ3 may be regarded as the first part, the second part and the third part of one tour λ_k for of the k^th ant, respectively, and form the tour λ_k.

Given the user pairs UP have been analogized as the (virtual) nodes, to apply the ACO algorithm to the resource allocation problem in the FD communication system, it is necessary to provide a proper analogy of the tour length L_k. In general, supposed that there are NRB resource blocks RB₁-RB_NRB to be allocated within the FD communication system, the virtual paths φ₁-φ_NRB form the virtual tour λ_k of the k^th ant, i.e., λ_k=φ1+ . . . +φ_NRB. Or, equivalently, the virtual tour λ_k includes the virtual paths φ₁-φ_NRB, and each virtual path φ_m includes a plurality of virtual edges e(x1_y1, x2_y2). Therefore, a virtual tour length Λ_k corresponding to the virtual tour λ_k may be expressed as

$\begin{array}{cc}\begin{array}{c}{\Lambda }_k=\mathrm{Length}\left({\lambda }_k\right)=\mathrm{Length}\left({\phi }_1\right)+\dots +\mathrm{Length}\left({\phi }_{\mathrm{NRB}}\right)\\ =d_k\left({\mathrm{RB}}_1\right)+\dots +d_k\left({\mathrm{RB}}_{\mathrm{NRB}}\right).\end{array}& \left(\mathrm{eq}.8\right)\end{array}$

In eq. 8, Length(⋅) represents a function outputting a virtual length/distance corresponding to a virtual tour or a virtual path, and d_k(RB_m) represents a virtual distance of the virtual path φ_m corresponding to the resource block RB_m. In an embodiment, the virtual distance d_k(RB_m) may be expressed as

$\begin{array}{cc}{d}_k\left({\mathrm{RB}}_m\right)=\sum _{\underset{x2\mathrm{\_y}2}{x1\mathrm{\_y}1}}{d}_{x1\mathrm{\_y}1,x2\mathrm{\_y}2}\left({\mathrm{RB}}_m\right),\forall {\mathrm{UP}}_{x1\mathrm{\_y}1},{\mathrm{UP}}_{x2\mathrm{\_y}2}\in {\phi }_m.& \left(\mathrm{eq}.9\right)\end{array}$

In eq. 9, the virtual edge distance d_{x1_y1,x2_y2}(RB_m) may be obtained according to eq. 5 or eq. 6, given that the resource block RB_m is assigned to the user pairs UP_{x1_y1}, UP_{x2_y2}, and UP_{x1_y1}, UP_{x2_y2}∈φ_m represents that the virtual nodes UP_{x1_y1}, UP_{x2_y2} are on the virtual path φ_m.

Given the virtual tour length Λ_k has been analogized as the actual tour length L_k, the modified ACO algorithm is presented in Table II, which is applied in the resource allocation problem in the FD communication system of the present disclosure.

TABLE II
1 // Initialization
2 Set t = 0
3 Set Nitr = 0
4 Set τi,j(t) = c for all edges and Δτi,j = 0 , ∀i∈   , ∀j∈   , ∀i,j∈   i≠j, , where   is the set
of nodes in cell b1,   is the set of nodes in cell b2,   is the set of BSs
5 while (Nitr< Nitr,MAX) or (Lk1 ≠ Lk2, ∀ k1, k2∈ {1,...,K}) do
6  for r = 1 to NRB do
7   Δt = 0
8   for step = 1 to NBS do
9    Δt = Δt +1
10    for k = 1 to K do
11     Define: UPi(r,step,k) : a virtual node which the kth ant stays at “step” of rth RB
12      if step = 1 and r ≠ 1 then
13       The kth ant chooses the virtual node UPi(r,step,k) as the starting point
14      else if step=1 and r=1
15       Uniformly Place the K ants on the (N•NBS) node to determine UPi(1,1,k)
16      else
17        The kth ant moves to the virtual node UPj(r,step,k) from virtual node
UPi(r,step,k) where UPi(r,step+1,k)=UPj(r,step,k)
18      end
19      Update Fk
20    end
21   end
22  end
23  for k = 1 to K do
24    Calculate Λk according to eq. 8
25    Update pheromonal trails according to eq.2, eq. 3 and eq. 4, where Lk in eq. 4 is
replaced with Λk
26    Update the shortest tour found
27  end
28 t = t + Δt
29 Nitr = Nitr + 1
30 Reset Δτi,j =0, ∀i∈   , ∀j∈   , ∀i,j∈   i≠j,
31 end
32 Print Shortest Tour

In Table II, the virtual node indices x_y have been abbreviated as the indices i or j. N, NBS, NRB represent a number of user pairs within one cell, a number of stations BS and a number of resource blocks RB. The ants presented in Table II are artificial ants, meaning that the processing unit 120 may simulate the behavior of the artificial ants. In some embodiments, N could be different in different cells.

On Line 10 of Table II, the k^th ant chooses to start from the virtual node UP_i(r,step,k) according to a plurality of initial probabilities p_I. In an embodiment, the initial probabilities p_I,i^k, corresponding to the virtual node UP_i for the k^th ant may be expressed as

$\begin{array}{cc}{p}_{1,i}^k=\{\begin{array}{c}\frac{\left[\mathrm{Rate}\left({\mathrm{UP}}_i\right)\right]}{\sum _{s\in \mathrm{Fk}}\left[\mathrm{Rate}\left({\mathrm{UP}}_s\right)\right]},\mathrm{if}i\in \mathrm{Fk}\\ 0,\mathrm{otherwise}\end{array}.& \left(\mathrm{eq}.10\right)\end{array}$

In an embodiment, the k^th ant may choose a virtual node based on the initial probabilities p_I,i^k.

On line 13, the k^th ant may choose the virtual node UP, based on the initial probabilities p_i,j^k(t). In some embodiments, line 13 can be replaced by line 15 in Table II.

On Line 17 of Table II, the k^th ant chooses to move to the virtual node UP, from virtual node UP_i according to the transition probability p_i,j^k (t) shown in eq. 1. On Line 25 of Table II, the virtual tour length Λ_k is substituted as the tour length L_k into eq.4, such that the virtual pheromonal trail τ_i,j(t) is updated according to eq. 2 and eq. 3. In an embodiment, the k^th ant may choose the virtual node UP_j based on the initial probabilities p_i,j^k(t).

Operations of the ACO algorithm applied in the resource allocation problem in the FD communication system may be summarized as a resource allocation method 60. The resource allocation method 60 may be compiled as the program code 124 to instruct the processing unit 120 to execute the resource allocation method 60.

In the resource allocation method 60, Step 602 may be referred to the computation of the virtual distances d_k(RB₁), . . . , d_k(RB_NRB) in eq. 9, which may be referred back to eq. 5 or eq. 6. Step 604 may be referred to the operation of Line 24 in Table II or the computation of eq. 8. Step 606 may be referred to the operation of Line 25 in Table II or the computation of eq.2, eq. 3 and eq. 4. Step 608 may be referred to the operation of Lines 8-22 in Table II, which has been illustrated in FIG. 4 and FIG. 5. In Step 610, the stopping criterion is presented on Line 5 of Table II. The resulting tour obtained in Step 612 may be referred to the shortest tour in Line 32 of Table II, in which the resulting tour λ* is obtained.

The resulting tour λ* includes virtual paths φ₁*-φ_NRB* corresponding to the resource blocks RB₁-RB_NRB respectively. In step 614, according to the resulting tour λ*, the resource block RB_m is allocated to the user pairs corresponding to the virtual nodes on the virtual path φ_m*.

Note that, since the virtual distance or the virtual tour length is inversely proportional to the data rate, or the virtual distance or the virtual tour length decreases as the data rate increases, the shortest tour would bring the largest data rate.

In summary, the present disclosure utilizes the ACO algorithm to perform multi-cell resource allocation, which maximizes the system data rate for the FD multi-cell communication systems.

Those skilled in the art will readily observe that numerous modifications and alterations of the device and method may be made while retaining the teachings of the invention. Accordingly, the above disclosure should be construed as limited only by the metes and bounds of the appended claims.

Claims

1. A resource allocation method, applied in a full duplex communication system comprising a plurality of stations and a plurality of user devices, wherein the full duplex communication system operates on a plurality of resource blocks, and the resource allocation method comprises the following steps: selecting a plurality of first selected virtual nodes according to a plurality of virtual pheromonal trails on a plurality of virtual edges, wherein the plurality of first selected virtual nodes form at least one virtual tour, and the virtual tour comprises a plurality of first virtual edges;updating the plurality of virtual pheromonal trails on the virtual edges according to virtual distances corresponding to the first virtual edges of the virtual tour;selecting a plurality of second selected virtual nodes according to the updated virtual pheromonal trails on the virtual edges, wherein the plurality of second selected virtual nodes forms at least one resulting virtual tour;allocating the plurality of resource blocks to a plurality of selected user pairs according to the resulting virtual tour, wherein the plurality of selected user pairs correspond to the resulting virtual tour;wherein the plurality of user devices and the plurality of stations forms a plurality of user pairs, and each user pair comprises a downlink user device and a uplink user device, among the plurality of user pairs, along with a station among the plurality of stations;wherein the plurality of user pairs represents a plurality of virtual nodes, and the plurality of first selected virtual nodes is among the plurality of virtual nodes;wherein a virtual edge, among the plurality of virtual edges, is formed between two virtual nodes among the plurality of virtual nodes;wherein the virtual distances are related to data rates.

2. The resource allocation method of claim 1, wherein the step of updating the plurality of virtual pheromonal trails on the virtual edges comprises: calculating the virtual distances corresponding to the first virtual edges of the virtual tour;determining incremental virtual pheromones corresponding to the first virtual edges of the virtual tour according to the virtual distances corresponding to the first virtual edges of the virtual tour; andupdating the plurality of virtual pheromonal trails on the virtual edges according to the incremental virtual pheromones.

3. The resource allocation method of claim 1, wherein a first virtual distance corresponding to a first virtual edge is inversely proportional to data rates of first virtual nodes between which the first virtual edge is formed.

4. The resource allocation method of claim 1, wherein a first virtual distance corresponding to a first virtual edge is inversely proportional to a data rate of one of first virtual nodes, in which the first virtual edge is formed between the first virtual nodes.

5. The resource allocation method of claim 1, wherein the step of selecting the first selected virtual nodes further comprises: forming a first feasible set comprising all or a part of the plurality of virtual nodes;selecting a third selected virtual node from the first feasible set, wherein the third selected virtual node is within a first cell;forming a second feasible set by removing a plurality of first virtual nodes within the first cell from the first feasible set, wherein the second feasible set comprises a plurality of second virtual nodes; andselecting a fourth selected virtual node from the second feasible set.

6. The resource allocation method of claim 5, wherein the step of selecting the third selected virtual node comprising: selecting the third selected virtual node according to a plurality of initial probabilities corresponding to the plurality of virtual nodes.

7. The resource allocation method of claim 6, wherein distribution of the initial probabilities is uniform distribution.

8. The resource allocation method of claim 5, further comprising: computing a plurality of transition probabilities corresponding to the plurality of second virtual nodes according to the plurality of virtual pheromonal trails on the plurality of virtual edges, wherein each transition probability is corresponding to a transition from the third selected virtual node to a second virtual node of the plurality of second virtual nodes; andselecting the fourth selected virtual node according to the plurality of transition probabilities.

9. A full duplex (FD) communication system comprising: a plurality of stations and a plurality of user devices, wherein the full duplex communication system operates on a plurality of resource blocks; anda computing device comprising a processing unit and a storage unit, wherein the storage unit is configured to store a program code to be executed by the processing unit to perform the following steps: selecting a plurality of first selected virtual nodes according to a plurality of virtual pheromonal trails on a plurality of virtual edges, wherein the plurality of first selected virtual nodes form at least one virtual tour, and the virtual tour comprises a plurality of first virtual edges;updating the plurality of virtual pheromonal trails on the virtual edges according to virtual distances corresponding to the first virtual edges of the virtual tour;selecting a plurality of second selected virtual nodes according to the updated virtual pheromonal trails on the virtual edges, wherein the plurality of second selected virtual nodes form at least one resulting virtual tourallocating the plurality of resource blocks to a plurality of selected user pairs according to the resulting virtual tour, wherein the plurality of selected user pairs correspond to the resulting virtual tour;wherein the plurality of user devices and the plurality of stations forms a plurality of user pairs, and each user pair comprises a downlink user device and a uplink user device, among the plurality of user pairs, along with a station among the plurality of stations;wherein the plurality of user pairs represents a plurality of virtual nodes, and the plurality of first selected virtual nodes is among the plurality of virtual nodes;wherein a virtual edge, among the plurality of virtual edges, is formed between two virtual nodes among the plurality of virtual nodes;wherein the virtual distances are related to data rates.

10. The FD communication system of claim 9, wherein the step of updating the plurality of virtual pheromonal trails on the virtual edges comprises: calculating the virtual distances corresponding to the first virtual edges of the virtual tour;determining incremental virtual pheromones corresponding to the first virtual edges of the virtual tour according to the virtual distances corresponding to the first virtual edges of the virtual tour; andupdating the plurality of virtual pheromonal trails on the virtual edges according to the incremental virtual pheromones.

11. The FD communication system of claim 9, wherein a first virtual distance corresponding to a first virtual edge is inversely proportional to data rates of first virtual nodes between which the first virtual edge is formed.

12. The FD communication system of claim 9, wherein a first virtual distance corresponding to a first virtual edge is inversely proportional to a data rate of one of first virtual nodes, in which the first virtual edge is formed between the first virtual nodes.

13. The FD communication system of claim 9, wherein the program code is executed by the processing unit to further perform the following steps: forming a first feasible set comprising all or a part of the plurality of virtual nodes;selecting a third selected virtual node from the first feasible set, wherein the third selected virtual node is within a first cell;forming a second feasible set by removing a plurality of first virtual nodes within the first cell from the first feasible set, wherein the second feasible set comprises a plurality of second virtual nodes; andselecting a fourth selected virtual node from the second feasible set.

14. The FD communication system of claim 13, wherein the step of selecting the third selected virtual node comprising: selecting the third selected virtual node according to a plurality of initial probabilities corresponding to the plurality of virtual nodes.

15. The FD communication system of claim 14, wherein distribution of the initial probabilities is uniform distribution.

16. The FD communication system of claim 13, wherein the program code is executed by the processing unit to further perform the following steps: computing a plurality of transition probabilities corresponding to the plurality of second virtual nodes according to the plurality of virtual pheromonal trails on the plurality of virtual edges, wherein each transition probability is corresponding to a transition from the third selected virtual node to a second virtual node of the plurality of second virtual nodes; andselecting the fourth selected virtual node according to the plurality of transition probabilities.