CN102096742A - Very large scale integrated circuit wiring design method based on taboo ant colony hybrid algorithm - Google Patents
Very large scale integrated circuit wiring design method based on taboo ant colony hybrid algorithm Download PDFInfo
- Publication number
- CN102096742A CN102096742A CN 201110043943 CN201110043943A CN102096742A CN 102096742 A CN102096742 A CN 102096742A CN 201110043943 CN201110043943 CN 201110043943 CN 201110043943 A CN201110043943 A CN 201110043943A CN 102096742 A CN102096742 A CN 102096742A
- Authority
- CN
- China
- Prior art keywords
- node
- taboo
- ant
- integrated circuit
- large scale
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a very large scale integrated circuit wiring design method based on a taboo ant colony hybrid algorithm. The method comprises the following steps of: placing n ants on a wiring source point, wherein the ith ant selects the next node to be reached from a starting node according to a state transfer rule; checking whether the starting node is an element in a taboo table which is configured to hold taboo elements; if information degree strength of the node is greater than a set maximum value, adding the node into the taboo table, otherwise, partially updating pheromones; adding the iterative nodes into the current route and recording a shortest route; and fully updating the pheromones according to a formula after all the ants finish the shortest routes. By the method, the wiring at both ends of a very large scale integrated circuit is described by using an ant colony system algorithm, so the distribution of initial pheromones is improved, proper parameter values are selected, the algorithm convergence time and the algorithm characteristics are brought in and used as restriction conditions, and a taboo search method is brought in; moreover, the method has high precision and is efficient.
Description
Technical field
The present invention relates to the wiring of VLSI (very large scale integrated circuit) (VLSI), refer in particular to the two ends Wiring design method of VLSI (very large scale integrated circuit).
Background technology
Expansion day by day along with the integrated circuit scale, traditional Wiring design method can not meet the demands, the space-time complexity of the algorithm collapse occurred, finding the solution is too high, can't be near problems such as optimum solutions, for addressing these problems, some have appearred in recent years based on intelligent algorithm, bionic Algorithm such as genetic algorithm, neural network, immune algorithm, the new Wiring design method of ant group system etc.The Wiring design method based on ant group system wherein has the following disadvantages: (1) needs long search time.At the initial stage that algorithm is carried out, because pheromones distribution spatially is uniform, this just means that searching for the ant group system that arrives optimum for the feedback that relies on pheromones is difficult to find at short notice a shortest path of the overall situation, this is finding the solution space performance more outstanding when big, ants have only by constantly on the way staying pheromones, As time goes on, can be by the pheromones on the many paths of ant quantity significantly more than other path, this process is very long relatively.(2) stagnation behavior appears in algorithm easily.Along with the continuous accumulation of pheromones on the path, after search proceeded to a certain degree, all are individual, and separating of finding was in full accord, can not further search for solution space, is unfavorable for finding globally optimal solution, promptly algorithm to a certain degree be absorbed in local optimum.
Summary of the invention
The objective of the invention is provides a kind of effectively based on the VLSI (very large scale integrated circuit) Wiring design method of taboo ant group hybrid algorithm for overcoming above-mentioned the deficiencies in the prior art, makes VLSI (very large scale integrated circuit) two ends wires design more efficiently reach low cost.
The technical solution used in the present invention comprises the steps: the strongly connected graph of (1) initialization VLSI (very large scale integrated circuit), determine the position of the initial smell of net point and the wiring source point start and the impact point end of strongly connected graph, the pheromones that covers simultaneously in this minimum rectangular area of 2 is strengthened;
(2) will
nAnt places wiring source point start, and i ant wherein is by the state transitions rule
The follow node
rSelect the next node that will arrive
s, and towards impact point end direction search;
SRAccording to the selective rule formula
Determine;
Ph(
s) be node
sOn pheromones intensity;
Dis(
R, s) be node
rTo node
sDistance;
Be the distance of source point start to impact point end;
а,
βIt is parameter;
qIt is [0,1] interval equally distributed random number;
q 0Be 0≤
q 0The parameter of≤1 scope:
Allowed(
s) be the set of selectable next point,
s∈
Allowed(
s); (3) check node
sWhether be for concluding the element in the taboo table that the taboo element is provided with, if node
sMaximal value that surpass to set of information degree intensity
Ph_max, then with this node
sAdd in the taboo table, otherwise, press
Carry out the plain renewal of local message, check again after the renewal;
Ph 0Be the plain intensity of initial information, ρ is the volatilization parameter of pheromones; (4) whether determine with node according to taboo table information
sList in the optional some set, till this ant finds impact point end maybe can select set of node to be sky; Otherwise, get back to step 2) undertaken by the state transitions rule, successively the circulation; (5) each the node s with iteration all joins current path and writes down shortest path, all finishes the shortest path back-pushed-type of oneself all ants
Carry out the plain renewal of global information, wherein, update coefficients
,
L Gb Be current global optimum path,
CIt is constant; (6) check whether all ants all set out and found impact point end, if then export shortest path, if not, then get back to step 2) ant restart from source point start.
The present invention uses ant group system performance, with ant group system algorithm the wiring of VLSI (very large scale integrated circuit) two ends is described, the distribution of improvement initial information element, select suitable parameter value, introduce algorithm convergence time and characteristic as constraint condition and introducing tabu search method, propose taboo ant group system, form a kind of new VLSI (very large scale integrated circuit) Wiring design method, have very high accuracy and high efficiency based on complementary hybrid intelligent algorithm.
Description of drawings
Fig. 1 is a schematic flow sheet of the present invention;
Fig. 2-the 3rd, the structural representation of the embodiment of the invention.
Embodiment
Referring to Fig. 1, the concrete implementation step of the present invention is as follows:
Strongly connected graph to VLSI (very large scale integrated circuit) carries out initialization earlier, determine the initial smell of strongly connected graph net point, determine the position of wiring source point start and impact point end on the strongly connected graph again, the pheromones in drape line source point start and this minimum rectangular area of 2 of impact point end is simultaneously strengthened.Then will
nAnt places wiring source point start,
nThe impact point of ant is end.
nThe current location that i ant in the ant sets out is a node
r, what this i ant selection next one will arrive is node
sConsideration adds the weight from the impact point distance in the state transitions rule, ant has direction ground to search for towards impact point.In conjunction with the tabu search method, prevent the local optimum that may occur, i ant selected the next node that will arrive
sThe state transitions rule be:
In the formula (1)
Ph(
s) be the node that the next one will arrive
sOn pheromones intensity.
Dis(
R, s) be the node of current location
rTo the next node that will arrive
sDistance.
Be the distance of source point start to impact point end.
а,
βBe parameter,
а,
βThe relative importance that has reflected heuristic information (pheromones intensity) and local path.
qIt is [0,1] interval equally distributed random number.
q 0Be a parameter (0≤
q 0≤ 1),
q 0Embodied the relative importance of utilizing between priori and the exploration new route.
SRBe the rule of determining according to following formula (2) of ratio at random:
In the formula (2)
Be to the next node that will arrive by the node r of current location
sSelective rule, promptly i ant selected the next node that will arrive according to the probability of formula (2)
S.Allowed(
s) be the set of the next one point that can select, can select set of node, and
s∈
Allowed(
s), short of obstacle, the node of current location
rPoint up and down can as the back reconnaissance, from
Allowed(
s) select than advantage according to rule again in the set.
By formula (1) as can be known, be positioned at the node of current location
rI ant selecting the next node that will arrive
sThe time, if random number
q≤
q 0, selecting down a bit according to formula (1), otherwise select down a bit according to formula (2), the node s that selects according to formula (2) rule adds current path.
Check selected node
sWhether be the element of avoiding in the Table A, if node
sMaximal value that surpass to set of information degree intensity
Ph_max, then with this node
sAdd in the taboo Table A.Avoiding Table A is for concluding the set that the taboo element is provided with, avoid Table A and begin to be empty set, surpassing the maximal value of setting when the pheromones intensity of certain node
Ph_maxAfter, this node can be joined in the taboo Table A automatically.If node
sInformation degree intensity surpass maximal value
Ph_max,Then according to following formula (3) to this node
sCarry out the plain renewal of local message:
Ph 0Be the plain intensity of initial information, ρ represents the volatilization parameter of pheromones.Ant is whenever by a node, when the information degree intensity of this node surpasses maximal value
Ph_maxThe time to carry out all that local message is plain to be upgraded to this node, local message is plain checks the plain intensity of this nodal information after upgrading again, if surpass maximal value
Ph_max, this node is joined the taboo Table A and upgrades the taboo Table A.To the node in the taboo Table A, whether this node is listed in the optional some set according to the decision of taboo Table A information.So, till this ant finds impact point end maybe can select set of node to be sky, if this ant does not find impact point end maybe can select set of node
Allowed(
s) be not sky, the node of selecting the next one to arrive by the state transitions rule of formula (1) again.
I ant find impact point end maybe can select set of node be till the sky after, i+1 the node that ant selects the next one to arrive by the state transitions rule of formula (1), circulation so successively is till n ant all found impact point end.Each node s of iteration is all joined current path and writes down shortest path, and so-called current path is the path that ant is passed through in once searching process, and so-called shortest path is the optimal path that finds in all iterative process.The shortest path of each iteration is added among the taboo table B, the set that taboo table B is provided with for the node elements of concluding on the current local optimum path that has searched, so-called local optimum path is the optimal path that finds in an iterative process.All finish the shortest path of oneself all ants after, carry out the plain renewal of global information by following formula (4),
Δ in the formula (4)
Ph(
s) be update coefficients, determine by following (5) formula:
L Gb Be current global optimum path,
CBe constant; Have only those pheromones intensity that belong to the node on the global optimum path just can be enhanced, the just decay of the plain intensity of other nodal information.
At last, check whether all ants has all been set out and has been found impact point end, if do not have, then gets back to the step that ant is restarted from source point start, if all ants have all set out and found impact point end, then exports shortest path.
Parameter in the above-mentioned formula can select the superior to select and configuration according to particular problem, can get
а=1,
β=-1;
=
,
n=20, ρ=0.1,
q 0=0.5, here
q 0Size determined to utilize priori and explored the relative importance of new route,
q 0Value excessive or too smallly all can directly have influence on result of experiment, when
q 0=0.1 o'clock, because pheromones can not accumulate well, priori can not well remain, and therefore, though can find optimum solution, algorithm can't be restrained.When
q 0=0.9 o'clock, because
q 0Value is excessive, and pheromones accumulation is too fast, causes subsequent path explored presenting inertia, and the algorithm premature convergence is absorbed in local optimum, can't find globally optimal solution, therefore, and can be suitable by selecting
q 0The distribution of parameter value improvement initial information element.
1 embodiment of the present invention below is provided.
Embodiment
Referring to Fig. 2, black shade is a barrier, and it both can be a function element, also can be cloth gauze.The start point is the wiring source point, and T (end) point is an impact point, and V is the zone of can connecting up.Implement as follows:
Step 1: strongly connected graph is carried out initialization, determine the initial smell of net point;
Step-2: determine wiring source point start and impact point T (end) position, the pheromones that can cover in this minimum rectangular area of 2 is strengthened;
Step 3:
nAnt places source point start;
Step 4: in the grid node of Fig. 3, seek next node according to formula (1) and (2);
Step 5: explore each paths according to the method for step 4, the barrier that gets around gray area is searched for, and upgrades content in the taboo Table A;
Step 6: the local message element to node upgrades according to formula (3);
Step 7: the plain renewal of global information is carried out in the local optimum path of being found, prevented the appearance of local optimum; Strengthen the pheromones intensity on the global optimum path, the pheromones intensity on the non-optimal path that weakens guarantees that the ant group searches for towards the direction of optimizing;
Step 8: find optimal path
Start to T (end)
Avoid the wiring problem that ant group system can realize Fig. 2 with the present invention, calculate by algorithm programming of the present invention, the statistical conditions of experimental result are summarized as follows table:
The statistics of experimental result
Algorithm | Completion rate | Optimal path length | The convergence number of times | Convergence time |
Taboo ant group system | 100% | 425(mm) | About 220 times | About 150 milliseconds |
The taboo ant group hybrid algorithm of the VLSI VLSI (very large scale integrated circuit) wiring that present embodiment proposed has very high reliability and operation efficiency.
Claims (3)
1. VLSI (very large scale integrated circuit) Wiring design method based on taboo ant group hybrid algorithm is characterized in that as follows:
The strongly connected graph of (1) initialization VLSI (very large scale integrated circuit) is determined the position of the initial smell of net point and wiring source point start and the impact point end of strongly connected graph the pheromones that covers simultaneously in this minimum rectangular area of 2 to be strengthened;
(2) will
nAnt places wiring source point start, and i ant wherein is by the state transitions rule
The follow node
rSelect the next node that will arrive
s, and towards impact point end direction search;
SRAccording to the selective rule formula
Determine;
Ph(
s) be node
sOn pheromones intensity;
Dis(
R, s) be node
rTo node
sDistance;
Be the distance of source point start to impact point end;
а,
βIt is parameter;
qIt is [0,1] interval equally distributed random number;
q 0Be 0≤
q 0The parameter of≤1 scope:
Allowed(
s) be the set of selectable next point,
s∈
Allowed(
s);
(3) check node
sWhether be for concluding the element in the taboo table that the taboo element is provided with, if node
sMaximal value that surpass to set of information degree intensity
Ph_max, then with this node
sAdd in the taboo table, otherwise, press
Carry out the plain renewal of local message, check again after the renewal;
Ph 0Be the plain intensity of initial information, ρ is the volatilization parameter of pheromones;
(4) whether determine with node according to taboo table information
sList in the optional some set, till this ant finds impact point end maybe can select set of node to be sky; Otherwise, get back to step 2) undertaken by the state transitions rule, successively the circulation;
(5) each the node s with iteration all joins current path and writes down shortest path, all finishes the shortest path back-pushed-type of oneself all ants
Carry out the plain renewal of global information, wherein, update coefficients
,
L Gb Be current global optimum path,
CIt is constant;
(6) check whether all ants all set out and found impact point end, if then export shortest path, if not, then get back to step 2) ant restart from source point start.
3. the VLSI (very large scale integrated circuit) Wiring design method based on taboo ant group hybrid algorithm according to claim 1 is characterized in that: in the step (3), ant is whenever by a node, when the information degree intensity of this node surpasses maximal value
Ph_maxThe time to carry out all that local message is plain to be upgraded to this node.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201110043943 CN102096742A (en) | 2011-02-24 | 2011-02-24 | Very large scale integrated circuit wiring design method based on taboo ant colony hybrid algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201110043943 CN102096742A (en) | 2011-02-24 | 2011-02-24 | Very large scale integrated circuit wiring design method based on taboo ant colony hybrid algorithm |
Publications (1)
Publication Number | Publication Date |
---|---|
CN102096742A true CN102096742A (en) | 2011-06-15 |
Family
ID=44129837
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 201110043943 Pending CN102096742A (en) | 2011-02-24 | 2011-02-24 | Very large scale integrated circuit wiring design method based on taboo ant colony hybrid algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102096742A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104331569A (en) * | 2014-11-13 | 2015-02-04 | 哈尔滨工业大学 | Small-delay fault testing channel selection method for large-scale integrated circuit based on selection of critical nodes and ant colony optimization algorithm |
CN106934173A (en) * | 2017-03-24 | 2017-07-07 | 哈尔滨工业大学 | Based on the digital microcurrent-controlled chip on-line testing method that TABU search is combined with Artificial Potential Field Method |
CN111866877A (en) * | 2020-06-11 | 2020-10-30 | 南京邮电大学 | 5G physical layer security authentication method based on memory |
WO2021169468A1 (en) * | 2020-02-26 | 2021-09-02 | 福州大学 | Method for constructing x-structure steiner tree by taking into consideration voltage slew rate |
CN114844823A (en) * | 2022-04-07 | 2022-08-02 | 桂林电子科技大学 | Method for generating shortest link with must-pass point directed ring by improving ACO algorithm |
CN116702694A (en) * | 2023-07-07 | 2023-09-05 | 成都电科星拓科技有限公司 | Printed circuit board two-end wiring method, medium and device based on ant colony algorithm |
CN116738928A (en) * | 2023-07-07 | 2023-09-12 | 成都电科星拓科技有限公司 | Printed circuit board parallel disconnecting and re-distributing method, medium and device |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1588381A (en) * | 2004-07-06 | 2005-03-02 | 清华大学 | Rectangular steiner tree method of super large size integrated circuit avoiding barrier |
CN1604094A (en) * | 2004-11-16 | 2005-04-06 | 清华大学 | Right angle wiring tree method for wire length optimized obstacle passing |
-
2011
- 2011-02-24 CN CN 201110043943 patent/CN102096742A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1588381A (en) * | 2004-07-06 | 2005-03-02 | 清华大学 | Rectangular steiner tree method of super large size integrated circuit avoiding barrier |
CN1604094A (en) * | 2004-11-16 | 2005-04-06 | 清华大学 | Right angle wiring tree method for wire length optimized obstacle passing |
Non-Patent Citations (1)
Title |
---|
《中国优秀硕士论文全文数据库》 20090930 邱巍 基于禁忌蚁群系统的VLSI线网布线研究 1-3 , 第2009年09期 2 * |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104331569B (en) * | 2014-11-13 | 2017-05-24 | 哈尔滨工业大学 | Small-delay fault testing channel selection method for large-scale integrated circuit based on selection of critical nodes and ant colony optimization algorithm |
CN104331569A (en) * | 2014-11-13 | 2015-02-04 | 哈尔滨工业大学 | Small-delay fault testing channel selection method for large-scale integrated circuit based on selection of critical nodes and ant colony optimization algorithm |
CN106934173A (en) * | 2017-03-24 | 2017-07-07 | 哈尔滨工业大学 | Based on the digital microcurrent-controlled chip on-line testing method that TABU search is combined with Artificial Potential Field Method |
CN106934173B (en) * | 2017-03-24 | 2020-05-12 | 哈尔滨工业大学 | Digital micro-fluidic chip online testing method based on combination of tabu search and artificial potential field method |
WO2021169468A1 (en) * | 2020-02-26 | 2021-09-02 | 福州大学 | Method for constructing x-structure steiner tree by taking into consideration voltage slew rate |
CN111866877B (en) * | 2020-06-11 | 2023-03-24 | 南京邮电大学 | 5G physical layer security authentication method based on memory |
CN111866877A (en) * | 2020-06-11 | 2020-10-30 | 南京邮电大学 | 5G physical layer security authentication method based on memory |
CN114844823A (en) * | 2022-04-07 | 2022-08-02 | 桂林电子科技大学 | Method for generating shortest link with must-pass point directed ring by improving ACO algorithm |
CN114844823B (en) * | 2022-04-07 | 2024-03-05 | 桂林电子科技大学 | Necessary point directed band loop shortest link generation method for improving ACO algorithm |
CN116702694A (en) * | 2023-07-07 | 2023-09-05 | 成都电科星拓科技有限公司 | Printed circuit board two-end wiring method, medium and device based on ant colony algorithm |
CN116738928A (en) * | 2023-07-07 | 2023-09-12 | 成都电科星拓科技有限公司 | Printed circuit board parallel disconnecting and re-distributing method, medium and device |
CN116738928B (en) * | 2023-07-07 | 2024-03-29 | 成都电科星拓科技有限公司 | Printed circuit board parallel disconnecting and re-distributing method, medium and device |
CN116702694B (en) * | 2023-07-07 | 2024-05-24 | 成都电科星拓科技有限公司 | Printed circuit board two-end wiring method, medium and device based on ant colony algorithm |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102096742A (en) | Very large scale integrated circuit wiring design method based on taboo ant colony hybrid algorithm | |
CN104617985B (en) | Power telecom network line optimization method and device based on ant group algorithm | |
CN102778229A (en) | Mobile Agent path planning method based on improved ant colony algorithm under unknown environment | |
CN109445444B (en) | Robot path generation method under barrier concentration environment | |
CN109102124B (en) | Dynamic multi-target multi-path induction method and system based on decomposition and storage medium | |
CN103648139A (en) | Cultural ant colony algorithm-based wireless sensor network node deployment design method | |
CN107357965A (en) | A kind of path planning design method of wind power plant collection electric line | |
CN110196602A (en) | The quick underwater robot three-dimensional path planning method of goal orientation centralized optimization | |
CN103473248A (en) | Continuously-moving K adjacent query method in road network environment | |
CN114020045B (en) | Unmanned aerial vehicle track planning method based on improved ant colony algorithm | |
CN108875898A (en) | Based on firefly-ant group algorithm low voltage power line carrier dynamic routing method | |
CN101867519B (en) | Dynamic area routing method and system for ad hoc network | |
CN107332770A (en) | One kind must be through a routed path system of selection | |
CN111385853B (en) | Directional diffusion routing method based on improved ant colony algorithm in wireless sensor network | |
CN103425753A (en) | Method for searching for heuristic shortest path based on direction optimization | |
CN101494599A (en) | Method for implementing multicast energy-saving routing protocol of Ad Hoc network based on layer-striding mechanism | |
CN109635998A (en) | A kind of adaptive Multipurpose Optimal Method solving vehicle routing problem with time windows | |
CN109902391A (en) | A kind of oil-gas pipeline planing method based on random volatility coefficient ant group algorithm | |
CN106656814A (en) | Routing method for power line carrier communication network | |
CN115696494A (en) | Large-scale ad hoc network multipoint relay selection method based on ant colony optimization | |
CN112866120A (en) | SDN flow table loop-free consistency updating method and system based on classified search | |
CN112000105A (en) | Mobile robot path planning method based on exchange strategy ant colony algorithm | |
CN114640619B (en) | Space-based network topology design method based on average communication distance and related equipment | |
CN116627175A (en) | Unmanned aerial vehicle path planning method based on improved ant colony algorithm | |
CN101216722A (en) | Clock management process and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C12 | Rejection of a patent application after its publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20110615 |